MEDICAL SCHOOL LHIBl&AI&lf Howard W. Estill Memorial ULTRAVIOLET RADIATION OTHER BOOKS BY THE SAME AUTHOR 1. Color and Its Applications, 1916, 1921. 2. Light and Shade and Their Applications, 1916. 3. The Lighting Art, 1917. 4. The Language of Color, 1918. 6. Artificial Light, Its Influence Upon Civilization , 192 6. Lighting The Home, 1920. 7. Visual Illusions, 1921. 8. The Book of The Sky, 1922. ULTRAVIOLET RADIATION ITS PROPERTIES, PRODUCTION, MEASUREMENT, AND APPLICATIONS BY M. LUCKIESH DIRECTOR OF APPLIED SCIENCE, NELA RESEARCH LABORATORIES NATIONAL LAMP WORKS OF GENERAL ELECTRIC CO. Author of "Color and Its Applications," "Light and Shade and Their Applications," "The Lighting Art," "The Language of Color," "Artificial Light Its Influence Upon Civilization," "Light- ing The Home," "Visual Illusions," "The Book of The Sky," etc. NEW YORK D. VAN NOSTRAND COMPANY EIGHT WARREN STREET 1922 COPYRIGHT, 1922 BY D. VAN NOSTRAND COMPANY Printed in the United States of America >lf Trinity College in Cambridge, stands a marble statue of Sir Isaac Newton holding a prism in his hand. This thoughtful silent face impressed Wordsworth as, "The marble index of a mind forever Voyaging through strange seas of Thought, alone." It is to the memory of this great man, who illuminated the pathway leading to the discovery of invisible radiation, that this book is dedicated. PREFACE During the six score years which have elapsed since the discovery of ultraviolet radiation a great deal of atten- tion has been given to its properties, production, and applications. As a consequence of this widening acquaint- ance with ultraviolet radiation, this form of energy is now of practical value to the chemist, the physicist, the engineer, the biologist, the ophthalmologist, the physician and others. Many sources of ultraviolet radiation are now available and the applications are rapidly increasing in number. Unfortunately much of the literature on the subject is confusing, owing to the lack of care in the choice of definitions, and limited in value, owing to care- lessness in specifying important factors such as those pertaining to the spectral character of the radiation. It is the primary aim of this book to present authentic data of such scope as to be useful to those who are interested in ultraviolet radiation. Theory has pur- posely been subordinated to experimental facts because the latter are not affected by the inevitable changes in theory. The author has drawn freely from the work of others although by no means is it claimed that all the best work has been included. In covering the scope in mind it has been necessary to choose among a large number of investigations. Many references have been presented and it is hoped that these will increase the usefulness of the book. March 20, 1922. M. LUCKIESH vii CONTENTS Chapter Page I. Introduction 1 II. Solar Radiation 15 III. Transparency of Gases 35 IV. Transparency of Liquids 46 V. Transparency of Solids 72 VI. Transparency of Glasses 79 VII. Reflection of Ultraviolet Radiation 93 VIII. Ultraviolet Radiation in Common Illuminants 107 IX. Experimental Sources 133 X. Detection and Measurement 165 XI. Effects Upon Living Matter 204 XII. Various Photochemical Effects.. 223 LIST OF PLATES PLATE PAGE I. Absorption Spectra of Eye-Media Frontispiece II. The transmission of various media for ultraviolet radiation from the iron arc as obtained by a quartz prism spectrograph 72 HI. Two ultraviolet spectra of the tungsten arc and two of the iron arc as obtained by a quartz prism spectrograph 93 IV. Ultraviolet spectra of the ordinary carbon arc, the iron arc, and the quartz mercury arc as obtained by various photographic ex- posures 107 V. The white flame arc a powerful source of "near" ultraviolet radiation 116 VI. The quartz mercury arc shown with the quartz arc exposed and also as used for exposing materials to its radiation 127 VII. The ultraviolet spectra of the tungsten arc through quartz, at various currents and photographic exposures 133 VIII. Ultraviolet transmission spectra of clear and cobalt glasses as ob- tained by a quartz prism spectrograph 165 IX. The radiation from the quartz mercury arc employed in a recir- culating drinking-water system in a large factory 204 X. Radiant energy is finding many applications in therapeutics 211 XI. The white flame arc as used in dye-testing 223 XII. The carbon arc as used in blue-printing 237 ULTRAVIOLET RADIATION CHAPTER I INTRODUCTION When Newton was an old man, acknowledged by civi- lization as the greatest of his race, he wondered what the world would think of his labors, adding, " It seems that I have been but as a child playing on the sea-shore, now finding some pebble rather more polished, and now some shell rather more agreeably variegated than an- other, while the immense ocean of truth extended itself unexplored before me." This statement is characteristically modest. It is true that a great ocean of truth remained to be explored, but Newton 1 discovered this ocean when, in the year 1666, he found his way to the " sea-shore." Before that time something was known of the laws of reflection, refrac- tion, and transmission of light, but Newton's epoch- making discovery of the spectrum revealed an unexplored region pertaining to the composition of light and of radiant energy in general. When he placed his glass prism near the hole which he had cut in a window shade, he expected to see the beam of light refracted and end on his vertical screen in beautiful colors. But this thinker was not content merely to perform an experiment with which he was familiar. He applied himself to inquiry pertaining to the " vivid and intense colours " as ex- pressed in his own words. This led to the discovery of the variation of refrangibility with the wave-lengths of the radiant energy or radiation. Newton saw only the visible radiation light and gave to the spectral colors, which he saw on the screen, ULTRAVIOLET RADIATION the names, violet, indigo, blue, green, yellow, orange, and red. It is now known that with suitable apparatus more than one hundred distinctly different hues may be readily distinguished in the visible spectrum. 2 Although New- ton throughout the remainder of his life made numerous excursions into that unexplored " ocean of truth " and brought many facts to light he died without knowing that there was an extensive series of radiations, differing physically only in wave-lengths and frequency of vibra- tion, of which visible light was only a very small part. When he saw the colored patch on the screen he had no means of learning that beyond the violet edge ultra- violet radiation was impinging upon the screen and be- yond the red edge infra-red radiation was present. A silver compound would have detected the former; a deli- cate thermometer would have announced the latter. If his prism could have produced a normal spectrum of all the rays in the beam of solar energy and if his eyes could have seen it in its entirety from the extreme ultraviolet end to the farthest infra-red, he would have seen a spec- trum along the entire length of the room. Now it is known that radiant energy of the same physical character- istics and differing only in wave-lengths and in frequency of vibration, includes not only visible light but ultraviolet and infra-red radiation, electric waves of a great range of wave-lengths and even X-rays. If a normal spectrum of this entire range of wave-lengths of radiant energy could be produced and the portion of it due to visible radiation light were one foot long, the entire spectrum would be several million miles in length. In the foregoing the terms " radiation " and " radiant energy " have been used. They will be used interchange- ably throughout these chapters to express a certain form of energy. The term " light " appears to be best restricted to mean the sensation produced by visible radiation. Since Maxwell developed the electromagnetic theory, INTRODUCTION radiation or radiant energy is commonly termed elec- tromagnetic energy. The velocity at which radiant energy is transmitted through a perfect vacuum or through interplanetary space is independent of the fre- quency or wave-length and is approximately 186,300 miles per second or 3 x 10 10 centimeters per second. By divid- ing the velocity by the wave-length, the frequency is obtained. Three terms and symbols are in common use for desig- nating wave-lengths. Their relations and magnitudes are as follows: Unit Symbol Millimeters Relative length Angstrom A one ten-millionth 1 Millimicron mu or U.LL one millionth 10 Micron one thousandth 10000 For example the approximate boundary between the visible and the ultraviolet regions of the spectrum is expressed in the three units as follows: 4000 A, 400mji, 0.4ji. This corresponds to a frequency of 75 x 10 13 per second. In the following chapters the millimicron will be used as the unit of wave-length because it appears to meet the requirements more satisfactorily than the other units. Its use obviates the inconvenience of the decimal point preceding the units. Instead of employing the symbol pijx as has been the general custom the symbol m|i has been chosen as being strictly correct. The nature of radiation or radiant energy is still in doubt. The fact that it is transmitted through a vacuum calls for the creation by the imagination of a carrier. The "ether" was invented for this purpose and has served for many years. Many phenomena of light indi- cate that radiant energy is propagated in the form of ULTRAVIOLET RADIATION wave motion and that the waves are transverse. How- ever, these details, intensely interesting as they are scientifically, are not important from the viewpoint of this book. If the hypothetical " ether " must be dis- carded and the " waves " remodelled, the production, properties, and applications of ultraviolet energy will not be altered. Before entering upon a discussion of ultraviolet energy the entire spectral range of radiant energy will be pre- sented. The divisions are more or less arbitrary with the exception of visible radiation and even the limits of the visible spectrum are not well defined. The terms applied to the various regions have developed from certain prop- erties or uses of radiations of the various ranges of wave-length. Sometimes this has caused confusion. For example, the terms, " chemical rays " and " actinic rays " are misleading. They have been applied to radiation in- cluding ultraviolet and that corresponding to the short- wave (blue and violet) end of the visible spectrum, not- withstanding the fact that radiant energy of many other wave-lengths produce chemical changes. This looseness in terminology has inhibited progress to some extent and has often cast uncertainty over work which has deserved a better fate. Another case is the term " light." This is commonly used in three senses: (1) to express visual sensation; (2) to express radiant energy of wave-lengths included only in the visible spectrum; (3) to express the radiant energy throughout the entire (visible and invisible) spectrum. There appears to be no need for using the term in the third sense. The terms "ultraviolet radiation" and " infra-red radiation " are satisfactory for the invisible regions. If the use is confined to the first two, confusion would not be entirely eliminated but it would be greatly reduced. The author prefers to use the term " light " to express visual sensation ; and the term " visible radiation " INTRODUCTION to express the radiant energy of only those wave-lengths capable of exciting visual sensation. These meanings will be adhered to throughout this book. The approximate wave-length limits of the various spectral regions of radiant energy in the order of increas- ing wave-length are as follows: Wave-length Ultraviolet radiation 0-390m/x Extreme region 0-200 Gamma rays (arbitrary limits) to O.Olmju Rb'ntgen rays (arbitrary limits) 0.01 to 60m/z Middle or intermediate region 200-300 Near region 300-390 Visible radiation 390-770m/x Violet 390-430 Blue 430-470 Blue-green 470-600 Green 600-630 Yellow-green 630-660 Yellow 560-690 Orange 690-620 Red 620-770 Infra-red radiation . 77-coju Near region . 77- 20 Infra-red photography to 1/i Fluorite prism to 10/i Rock salt prism to 20/i Intermediate region 20-600 Selective reflection from rock salt to 50/x Selective reflection from potassium chloride to 61/z " Restrallen " method up to 354/z Electric oscillator method up to 600ju Extreme region 600-oo Electric waves such as that radiant energy studied by Herz, that used in wireless electric circuits, that resulting from high frequency currents, and that due to ordinary alternating currents complete a long range of wave-lengths to beyond 12 km. in wave-length. For convenience in description the ultraviolet and infra- red have been divided into three regions, namely, extreme, middle (or intermediate), and near (in reference to the ULTRAVIOLET RADIATION visible region). This has been found to be convenient for descriptive purposes. The practical limits of the visible spectrum from the standpoint of the eye are 400m^ and 700m[i. The approximate wave-length limits of the principal spectral hues are presented because of the con- venience of such terms as " green," " blue," etc. How- ever, when such terms are used for descriptive purposes the user should be certain that they actually represent approximately the spectral range. A spectroscope will determine this with certainty. It should be thoroughly realized that the eye is synthetic and not analytical. For example, a blue may appear blue but when examined by means of a spectroscope it may be seen to have a red band. Another striking case is yellow. It may appear yellow to the eye and still consist only of red and green. Yellow filters are commonly assumed to absorb ultra- violet radiation but many of them transmit some of the near ultraviolet. These are only three of many errors which can be easily made if the transmitted radiation is not analyzed by an analytical instrument. 2 On considering the spectrum of radiant energy as a whole and the various properties of the radiation of various spectral ranges it is seen that visible radiation is merely that of a certain range of wave-lengths to which the visual sense responds. A wireless receiving station in an analogous manner is tuned to respond to radiant energy of a certain range of wave-lengths or frequencies. Silver chloride is affected by radiation of a certain range of frequences corresponding to ultraviolet and violet. Each of the various photographic emulsions " sees " a certain range of wave-lengths of radiation. This chem- ical process responds to certain waves and that to others and so on. Thus viewed as a whole, if we may stretch the analogy, there are a great many " eyes " varying in sensibility to radiant energy of various ranges of wave- length or frequency. INTRODUCTION It is unfortunate that more accurate data have not been given in many investigations especially pertaining to the wave-lengths of radiant energy employed. Too often such indefinite terms as " ultraviolet light," " actinic rays," " blue light," etc., have been used in describing results. In other words, the spectral limits and the spec- tral distributions of energy have not been determined. Many of the effects of radiation upon plant-life, in indus- trial chemistry, in therapeutics, and in other fields are still uncertain because of the lack of specifications as to spectral limits and energy-distribution. Little definite progress can be made without a spectro- graph and its optical system must be of quartz or of other material transparent to the near and middle ultraviolet regions. Even quartz fails for work in the extreme re- gion. It should usually be possible to ascertain the spec- tral limits and energy-distribution of the illuminant used. Knowing these and especially the former, it is possible through characteristics of various substances, to define at least approximately, the radiation employed. Often there is no question regarding the spectral limits of the radi- ation of chief interest but there may be doubt as to the presence of other radiation. For example, effects are sometimes attributed to ultraviolet radiation when visible radiation is present. What effect may the latter have had? Another common example is the use of a blue glass with the assumption that blue light is the only radi- ation present. Most blue glasses transmit some of the near ultraviolet and the near infra-red. The commonest blue glass (cobalt) generally also transmits deep red and some infra-red radiation. Such looseness has caused a great deal of confusion and is responsible for the chaotic state of various phases of the use of radiant energy. Although it was not until after Newton's discovery of the visible spectrum that analytical experiments were performed regarding the photo-chemical properties of 8 ULTRAVIOLET RADIATION light, the observations of the keen intellects of preceding centuries should not be overlooked. The part played by daylight in the production of the green coloring matter in plants was observed long before the Christian era. Aristotle noted this influence as early as the year 350 B.C. The bleaching of pigments and of other coloring media was also known in those early centuries. Even the alchemists hinted of knowledge of the existence of such photo-chemical effects. However, it was not until the latter half of the 17th century that really scientific observations were made. Throughout the 18th century many facts were garnered from the unknown which paved the way for the rapid progress in the 19th century to which men like Bunsen and Roscoe contributed so much. It is not the aim to present a detailed and chronological account of the development of the science of photo- chemistry, but rather to provide a hurried glimpse of the course of knowledge pertaining to effects of radiant energy in order that the reader may better appraise the status of our knowledge at the present time. Ten years after Newton described his decomposition of sunlight into its component colors by means of the prism, and dis- covered the variation of refrangibility with the hue, Romer, a Danish astronomer, discovered that light travelled at a finite velocity. It is remarkable that his determination of this velocity differed by only three per cent from the value accepted at the present time. Huy- gens enunciated the wave-theory of light in 1678. All these views must be modified to harmonize with present knowledge but this does not depreciate the value of the work of the pioneers. Science is ever in a state of flux and it is not expected that theories of today will still be unmodified tomorrow. While dealing with radiation in these chapters scientists are busy undermining prevalent ideas of continuous propagation of radiation, the wave- theory of the past, the hypothetical ether, and other INTRODUCTION " models " upon which scientific progress is built. But there is consolation in the thought that experimental facts are not altered. It is the interpretations and explanations that suffer. Scheele in 1777 projected the visible spectrum upon silver chloride and noted the release of the chlorine and the production of metallic silver in the region of the violet rays. He was on the verge of discovering the ultraviolet region but it escaped his attention. However, he ob- served many photo-chemical reactions. The infra-red and ultraviolet regions were discovered almost simul- taneously. W. Herschel announced the former in 1800 and Ritter in 1801 noted the effect on silver chloride of what proved to be ultraviolet radiation. In the succeeding years several discoveries were made pertaining to the effect of visible and ultraviolet radiation upon silver salts, especially silver chloride. Gay Lussac and Thenard in 1809 noted various effects of light on chlorine and hydrogen. Already such chemical effects as the bleaching of chlorophyll and various dyes, and the decomposition of water by chlorine in bright sunlight had become known. In 1815 Planche noted the effects of light upon many metallic salts and a few years later Grot- thus enunciated this photo-chemical absorption law, " Only the rays absorbed are effective in producing chem- ical change." Discoveries of photo-chemical effects followed one after the other very rapidly in the early part of the nineteenth century. Chevreul, a pioneer in the science of color, described in 1837 the influence of air and moisture, in con- junction with sunlight, in the bleaching of vegetable colors. Although Scheele, Ritter and others paved the way for the development of photography, Niepce and Daguerre produced the first practicable process in about 1830. E. Bequerel contributed much to photo-chemistry and many analytical researches in connection with the effect 10 ULTRAVIOLET RADIATION of radiation of various wave-lengths on silver salts. Grad- ually the science of photography developed until in 1873 Vogel increased the spectral range of sensitiveness of silver salts by introducing certain dyes. These are only a few of the highlights of the evolution of photography. Since Vogel's time thousands of researches have devel- oped photography to a point where it is one of the best tools with which to invade the very realm whence it comes. Hurter and Driffield 3 during the latter portion of the past century extensively investigated the photo- graphic process. Maxwell in about 1868 enunciated the electromagnetic theory of radiation which among other things predicted the existence of electric waves. Herz in 1888 verified this as to certain electric waves much greater in wave-length than the longest infra-red waves which had been meas- ured. This left an unexplored gap in the long-wave region which has gradually been shortened in recent years until at the present time it may be said that this region has been almost completely explored. In 1895 Rontgen discovered the marvelous X-rays which later were classified as radiations of extremely short wave-lengths thus leaving an unexplored gap be- tween them and the shortest known ultraviolet radiations. This gap has been gradually shortened during the years which have elapsed since the discovery of X-rays. The limit of transparency of optical systems of glass is in general at about 340m^. As long as glass prisms were used for dispersing radiation into its spectrum the known spectrum could not be extended beyond the near ultraviolet. Quartz crystals were found to be transparent throughout the near and middle ultraviolet regions and in fact as far as ISSmjj,. lustruments employing quartz made it possible greatly to extend the known ultraviolet spectrum and the development of the art of fusing pow- dered quartz has further contributed to the accomplish- INTRODUCTION 11 ments in this field. The transparency of fluorite extends much further into the ultraviolet than that of quartz and by using this substance Schumann 4 during the period of 1890 to 1903 extended the explored region from 200 to The reflection-grating spectro graph which eliminates the necessity of employing transparent media in its con- struction now became a valuable accessory in the explora- tion of the extreme ultraviolet. It has been found that its reflection-factors for radiations of the extreme ultra- violet are sufficiently great for use in this work. How- ever, gases absorb quite strongly the ultraviolet radiation in this region and therefore the grating spectrograph has been enclosed in a compartment which may be evacuated. Gelatine is opaque to the extreme ultraviolet and inas- much as photography has been the chief means of record- ing these short-wave spectra it was necessary for Schu- mann to develop a special photographic plate. This plate was flowed with an emulsion containing the least amount of gelatine which could be used in its preparation. By eliminating the fluorite window it was possible to extend the explorations even to shorter wave-lengths. Lyman placed the light-source such as a spark in the spectrograph chamber and was soon able to extend the known spectrum to about 50mjx. The disruptive dis- charge or high potential spark has been a popular source of radiation for investigating this region. Saunders, Merton, McLennan and others have investigated the extreme ultraviolet between the region where Schumann's labors ceased and about 50m[i, but Lyman 5 may be con- sidered to be the pioneer. Thus it has been seen that the gap between the X-ray and the ultraviolet spectra has been greatly shortened during the score of years succeed- ing Rontgen's discovery. Recently it appears that Millikan 6 has completely spanned the gap, for apparently he has produced X-rays by 12 ULTRAVIOLET RADIATION means of a very high potential spark which he has re- corded by means of the photographic plate of the Schu- mann type. The remarkable success of Millikan and his colleagues in extending measurements to about 20mji was due chiefly to improvements in various directions. Michelson made for them special concave gratings which gave relatively more intense first-order spectra than those employed previously by others. They employed very high potential sparks in their evacuated spectrograph and kept the pressure in the latter below 10 ~~ 4 mm. Millikan concluded that his spectrograms recorded certain lines belonging to the X-ray spectrum of carbon, thus complet- ing the exploration of the gap between X-rays and known ultraviolet radiation of shortest wave-length. Until the advent of the electric dynamo the only ade- quate source of ultraviolet radiation was the sun. This is unsteady, uncertain, and discontinuous as a source and limited in spectral range in the ultraviolet region. The carbon arc was the first artificial source of appreciable powerfulness, but in the earlier years of the electrical age the production of ultraviolet radiation in this manner was costly. As the sciences of electricity and of light-pro- duction advanced, richer and more powerful sources of ultraviolet radiation appeared. However, it may be said that the 20th century was the first to see the advent of artificial sources of this energy sufficiently efficient and adequate to draw marked attention to applications on a commercial scale. Thus it is seen that the age of ultraviolet radiation of wide application has only recently dawned. It has found mankind in the possession of a great deal of general knowledge pertaining to the effects of radiant energy but somewhat lacking in specific details especially those pertaining to the spectrum. Technical literature contains thousands of references to effects of ultraviolet radiation but in many of them valuable specific data are lacking. INTRODUCTION 13 Perhaps the most general weakness is the absence of in- formation pertaining to the spectral limits and spectral distribution of energy. At least, if the source of the radiation is fully described and the spectral limits are specified there is much less doubt than in attributing a result merely to " ultraviolet energy." It is difficult to measure the intensity of ultraviolet radiation but it may be accomplished in several ways. Another difficulty is the absence of a continuous-spectrum source but a small quartz spectrograph will accomplish much in clarifying some points such as spectral limits and approximate quantitative spectral distributions of energy. Until the spectral aspects are given closer attention, progress in our knowledge of the effects of radiation will be slow and uncertain. The spectral transmission and reflection characteristic of many substances and the ultraviolet spectra of radia- tions from various sources are known to some extent. It is easy to obtain qualitative results of this nature by means of a quartz or reflection-grating spectrograph over a great range of the ultraviolet spectrum. Ordinary glass is opaque beyond the " near " region and is generally fairly opaque to energy of wave-lengths shorter than 340mji. Quartz is transparent to the "near" and " middle " regions, that is, down to the neighborhood of 185m|i. The extreme ultraviolet is easily absorbed by most known substances but it can be studied by means of a vacuum spectrograph. The spectral limits of the trans- mission characteristics of various media are discussed in other chapters. It is the aim here to emphasize the in- creasing difficulty of dealing with ultraviolet radiation as the wave-length decreases. In other chapters practicable methods and useful data are presented. 14 ULTRAVIOLET RADIATION References 1. Phil. Trans. (Abridged), Roy. Soc. Vol. i. 2. M. Luckiesh, Color and Its Applications, 1915, 1921. 3. Photographic Researches, Memorial Volume, 1920. 4. Ber. Wien. Akad., 102, Ila, 625. Smithsonian Contri- butions, 29, No. 1413, 1903. 5. Theodore Lyman, Spectroscopy of the Ultraviolet, 1914: Astrophys. Jour., 23, 1906, 181; 25, 1907, 45; 28, 1908, 52; 33, 1911, 98; 35, 1912, 341; 38, 1913, 282. 6. Astrophys. Jour. 52, 1920, 47; 53, 1921, 150. CHAPTER II SOLAR RADIATION Inasmuch as daylight is a very important factor in many chemical reactions and it is more or less a standard in many respects it appears necessary to discuss its characteristics in detail. For many years photography was almost solely dependent upon daylight but in recent years artificial illuminants have wrested supremacy in this respect from daylight. The testing of dyes and paints has been solely dependent upon sunlight until recently. Many other activities have been intimately associated with daylight but during the last score of years great strides have been made toward independence from daylight. In most cases where ultraviolet radiation is useful, results are desired without regard to their similarity to those obtained under daylight, but there are some cases, such as the testing of paints and dyes for permanency, where results similar to those obtained under daylight are of interest. For example, under ordinary conditions the foe of paints and dyes is daylight, for artificial illumi- nation is rarely sufficiently intense to have a marked effect upon their permanency. It is advantageous to test these materials under artificial radiation which is constant in intensity and controllable in every respect. However, the artificial radiation should yield results quite similar to that of daylight. This can be predicted if intensity and spectral character of the active radiation are approxi- mately the same for the artificial and natural radiations. Natural radiation, commonly called daylight, consists of (1) direct solar radiation, (2) diffuse radiation from the 16 ULTRAVIOLET RADIATION sky, and (3) radiation reflected from surroundings such as trees, buildings, etc. Radiation reflected from the surroundings is considerably modified owing to the selec- tive reflection of the various surfaces. In general ultra- violet radiation is materially reduced in quantity and in spectral range by reflection. These three classes of natural radiation vary in proportions over wide ranges. On overcast days direct solar radiation is reduced to zero and the radiation from the sky is greatly modified by the clouds. At noon on very clear days, the total light reach- ing the upper side of a horizontal surface when the entire sky is unobstructed, consists of clear blue skylight and direct sunlight. In such cases the skylight is about 10 to 20 per cent of the total, the latter percentage being common on average clear days. Noon sunlight is fairly constant in spectral character but its intensity varies with its altitude and with the condition of the atmosphere. * Therefore solar radiation varies in photochemical action momentarily, daily, sea- sonally and geographically. On clear days in mid- summer in the United States the intensity of illumination on a horizontal surface outdoors at noon reaches a value as high as 10,000 foot-candles. In midwinter on a clear day at noon the value is often only one-fourth or one-fifth as great for regions in the vicinity of 40 degrees latitude. On cloudy days, of course, the intensity of illumination may reach very low values but it is usually above 1000 foot-candles during midday. It may be stated that the intensity of daylight outdoors for several hours during midday is measured in thousands of foot-candles. Ordi- nary intensities of artificial illumination are a few foot- candles. In other words unobstructed daylight during midday is commonly of the order of magnitude of a thou- sand times greater in intensity than that of ordinary arti- ficial illumination. By using very powerful artificial sources at short distances the intensities of daylight may be approached or even exceeded in some cases. SOLAR RADIATION 17 The sky is brightest when it is hazy or when a thin film of cloud is present but this brightness is obtained at the expense of direct solar radiation. It has been seen that the intensity of daylight outdoors can be measured in thousands of foot-candles. Indoors this is ordinarily re- duced to less than a hundred foot-candles excepting near some unobstructed skylights. Furthermore the spectral character of natural radiation arriving indoors is altered by selective reflection, by surroundings and by selective absorption by the glass of skylights. This is easily shown by taking a photograph of the spectrum of the radiation from the sun or the sky through an open window by means of a quartz spectrograph. Then close the window and take another spectrogram. It will be seen that the latter does not extend quite as far into the ultraviolet as the former. Therefore where the greatest intensity of ultraviolet radiation is desired the unobstructed roof is the best place for operations. High altitudes offer advan- tages by escaping from much of the absorption of ultra- violet radiation by the lower atmosphere. At high alti- tudes such as attained by aircraft the relative clearness of the atmosphere is apparent by the extreme darkness of the sky and the relative greater brightness of the moon. The intensity of daylight at midday on a clear day is about 1,000,000 times greater than that due to the full moon at zenith. At noon the moon, viewed from the earth's surface, is about as bright as an average clear sky. The moon's disk is about J degree in diameter and there- fore occupies about 0.00001 of the total visible sky. Of course the moon's brightness is augmented by the bright- ness of the sky and therefore the moon cannot appear darker than the sky ; however, it is seen from the foregoing that the illumination due to average clear sky during mid- day is at least 100,000 times greater than that due solely to the full moon. Considering that the clear sky contributes only about one-fifth the total daylight at midday it is seen ULTRAVIOLET RADIATION that the intensity of illumination at noon on a clear day is of the order of magnitude of 500,000 times greater than that due to the full moon. If the former is taken as 5000 foot-candles the intensity of ilumination due to a full-moon at the zenith would be about 0.01 foot-candle. This is the order of magnitude. These magnitudes must be appreciated in many applications of ultraviolet radiation. It may also be of interest to note the rate at which energy is delivered to the earth by the sun. The solar constant as determined by the Astrophysical Observatory of the Smithsonian Institution from 696 observations over a period of ten years is 1.932 calories per minute upon a projected area of one square centimeter, that is, for normal incidence on a square centimeter per minute. Over a great circle of the earth with a circumference of about 25,000 miles (the projected area presented to the sun) solar radiation is received at the rate of about 2.3 x 10 15 horse-power. This equals an average rate of about 7000 horse-power per acre; about 1.45 horse-power per square yard; about 0.16 horse-power per square foot. To utilize this tremendous quantity of radiant energy is one of the problems to be solved. TABLE I Duration of Sunshine Latitude Degrees December 22 June 21 12 h 07 m 12 h 07 m 10 11 32 12 43 20 10 55 13 20 30 10 13 14 05 40 9 19 15 01 50 8 04 16 23 60 5 62 18 52 65 3 34 22 03 SOLAR RADIATION The duration of sunshine is of importance in many respects but for industrial processes solar radiation is not generally powerful enough throughout the entire period between sunrise and sunset. Table I gives some values of interest but more extensive data will be found else- where. 2 The percentage of cloudiness varies considerably, depending upon the season and the geographical location so that it would be futile to attempt to present adequate data here. In Table II the distributions of energy in the normal spectra of the radiation from the sun and from the sky are presented in arbitrary units as obtained by Abbott and his colleagues. The arbitrary unit is not of the same value for the last two columns. TABLE II Distributions of Energy in Radiation from Sun and Sky at Mt. Wilson Arbitrary units Wave-length 1 , m/x Sun Sky 422 186 1194 457 232 986 491 227 701 666 211 395 > 614 191 231 660 166 174 The maximum for solar radiation in Table II is in the region of 470mji but this is for relatively clear air at an altitude of 1730 meters or 5700 feet. At lower altitudes the sunlight is yellower, that is the maximum of the spec- tral energy-distribution shifts toward longer wave- lengths. (See Table V.) The spectral distributions of energy in solar and in sky radiation may be found else- where 3 compared with those of other illuminants. The 20 ULTRAVIOLET RADIATION solar spectrum extends to the neighborhood of 290mji and although it diminishes in intensity in the ultraviolet region it really ends rather abruptly indicating a powerful absorption-band in the earth's or the sun's atmosphere. Dry atmosphere in general selectively scatters radiation of the shorter wave-lengths. The selectivity varies con- siderably with atmospheric conditions but for dry air (the layer vertically above Mt. Wilson, altitude 1730 meters and barometric pressure 620mm.) the spectral transmis- sion-factors as obtained by Fowle 4 are presented in Table III. TABLE m Spectral Transmission-Factors of Dry Atmosphere Above Mt. Wilson m/i Per cent m/z Per cent 360 66.0 674 90.5 384 71.3 624 92.9 413 78.3 653 93.8 452 84.0 720 97.0 603 88.5 986 98.6 635 89.8 1740 99.0 The values in Table III agree very well with those expected from purely molecular scattering. When mois- ture and dust are present in the atmosphere the spectral transmission-factors are lower in value and are altered somewhat with respect to each other but there is still the same general increasing absorption with decreasing wave- length. The transmission-coefficients of the atmosphere as com- puted from the equation e m = e a m are given in Table IV for several places of observation (Washington, Mt. Wilson, Mt. Whitney) according to SOLAR RADIATION 21 the annals of the Astrophysical Observatory. In the foregoing equation e m is the intensity of the solar radiation after transmission through a mass of air, m; m is unity when the sun is at the zenith ; e is the energy which would have reached the point of observation if the atmosphere were perfectly transparent; a is the fractional amount, e m /e , actually reaching the point of observation when the sun is at the zenith. TABLE IV Transmission of Atmosphere. Transmission Coefficient, a mji Washington Mount Wilson Mount Whitney One mile nearer earth 300 0.460 0.550 320 0.520 0.615 340 0.580 0.692 360 0.635 0.741 380 0.380 0.676 0.784 0.562 400 0.560 0.729 0.809 0.768 460 0.690 0.832 0.887 0.829 500 0.733 0.862 0.919 0.850 600 0.779 0.900 0.940 0.866 700 0.858 0.950 0.964 0.903 800 0.886 0.970 0.976 0.915 1000 0.922 0.980 0.975 0.941 1600 0.938 0.976 0.965 0.961 2000 0.912 0.970 0.932 0.940 Nearer sea-level the values are smaller than those in Table IV. This is indicated in the data because the point of observation was lowest for Washington, next for Mt. Wilson, and highest for Mt. Whitney. Similarly the values computed for a point one mile lower than the ob- servatory on Mt. Whitney indicates the rapid decrease in the intensity of radiation of the shorter wave-lengths. The foregoing are for zenith depths of atmosphere. As the sun declines or decreases in altitude the air-mass 22 ULTRAVIOLET RADIATION increases approximately as the secant of the zenith dis- tance. Thus if m is unity for zenith sun, m equals 2 when the sun is at 60 degrees. At 70 degrees m equals about 2.9; at 80 degrees m equals about 5.6; at 88 degrees m equals probably about 20. The effect of air-mass upon the intensity and spectral distribution of solar radiation is shown in Table V. Any column may be plotted to obtain the spectral distribution curve of solar radiation. TABLE V Relative Spectral Values of Solar Radiation for Various Air-masses Washington Mt. Wilson Mt. Whitney mp. m = m = 1 m = 2 m = l m = 4 m = l 300 54 25 2 30 320 111 58 8 68 340 232 135 26 160 360 302 192 49 224 380 354 134 51 239 74 278 400 414 232 130 302 117 335 460 618 426 294 514 296 548 600 606 441 323 622 334 557 600 504 393 306 454 331 474 700 364 312 268 346 297 351 800 266 236 209 258 235 260 1000 166 153 141 163 154 162 The apparent black-body temperature of the sun as determined by the spectral distribution of solar radiation above the earth's atmosphere is between 6000 and 7000 degrees absolute (Centigrade + 273 deg.). At sea-level the apparent black-body temperature is between 5000 and 6000 degrees absolute owing to the reddening influence of the atmosphere. SOLAR RADIATION 23 The energy in the total solar spectrum may be con- sidered to be about equally divided between the infra-red and the visible (plus ultraviolet). Langley 5 obtained measurements indicating that about 60 per cent of the sun's radiation which reached the earth was in the infra- red region, the remaining 40 per cent being visible and ultraviolet. It is seen from Table V that most of the 40 per cent is in the visible region although a very appre- ciable amount of solar energy is in the region between 300 mpi and 400 m^. For equal amounts of total energy there is a greater proportion of " near " ultraviolet energy in sky than in solar radiation. The Fraunhofer lines in the solar spectrum are useful as comparison standards for spectroscopy although it is now usually more convenient to use the quartz mercury arc, a helium tube or some other source depending upon the region to be investigated. The principal Fraunhofer lines in the short-wave end of the solar spectrum are given in Table VI with the symbols sometimes used in desig- nating them and the elements responsible for the lines. TABLE VI Wave-lengths (Angstrom units) of some Fraunhofer Lines in the Solar Spectrum Symbol Element Wave-length Symbol Element Wave-length U Fe 2947.99 N Fe 3581.349 t Fe 2994.53 M Fe 3727.778 T Fe 3020.76 L Fe 3820.586 s Fe 3047.725 K Ca 3933.825 /* Fe 3100.046 H Ca 3968.625 \Si Fe 3100.430 h H 4102.000 Fe 3100.787 g Ca 4226.904 R Ca 3179.453 G /Ca 4307.907 Ca 3181.387 \Fe 4308.081 Q Fe 3286.898 f Fe 4325.939 P Fe 3361.327 G' or HT H 4340.634 Fe 3441.155 d Fe 4383.721 24 ULTRAVIOLET RADIATION The solar spectrum obtained by means of a quartz or reflection-grating spectrograph decreases very suddenly near 300 mji and ends abruptly near that wave-length. In this respect the radiation from electric incandescent lamps with glass bulbs approximates solar radiation because ordinary thin glass does not become totally opaque to ultraviolet radiation until the wave-length at which solar radiation disappears is approached. Of course, the spec- tral distributions are quite different but not to such an extent that many of the effects of solar radiation can be obtained with incandescent lamps when the intensity of illumination is of the same order of magnitude. Cornu 6 was one of the earliest to investigate the limit of the solar spectrum. He and his predecessors thought that the short-wave limit was due to selective absorption by the atmosphere. Therefore, in order to obtain the effect of variation in air-mass he photographed the sun's spectrum for various altitudes of the sun with the results indicated in Table VII. TABLE vn Limit of Solar Spectrum Time of day Limit in mju 10:30 295.5 0:02 295.0 1:18 295.5 1:50 297.0 3:09 299.0 3:40 304.5 4:17 304.5 4:38 307.0 The shortening of the spectrum with increasing air- mass naturally led him to think he had confirmed the assumption that the limit was due to ordinary atmos- SOLAR RADIATION 25 pheric absorption. From such data he developed a theory relating depth of air and the limiting wave-length. 7 His conclusion was that the spectrum would be extended 1 m^i if the photographs were made at an additional eleva- tion of 663 meters. His first altitude was 170 meters. He tested the accuracy of his formula by obtaining spec- trograms at altitudes of 660 meters and 2570 meters. His new results did not quite agree with his computations so he determined new constants, but his formulae will not be presented, for they were based upon the assumption that the limit of the solar spectrum was determined by the absorption of a homogeneous atmosphere whose density was distributed as indicated by the barometer. Cornu's formula certainly does not hold for the region from 180 to 240 mji. Among other conclusions Cornu concluded that the absorption of ultraviolet radiation by the atmosphere was due chiefly to nitrogen and to oxygen but the probable error of this conclusion will be seen later. He decided that solar radiation was a disinfectant chiefly upon the surface of bodies, apparently assuming that the active radiation was absorbed generally by thin layers of sub- stances. He further decided that any medium which absorbs the blue, violet, and ultraviolet rays of solar radia- tion such as glass, dust, fog, and clouds, inhibits the dis- infection desired from the standpoint of hygiene and sanitation. It is certain that consecutive days of rain and fog enhance the growth of pathogenic organisms but the active rays are not entirely absorbed by moderate depths of cloud and some clear glasses. Miethe and Lehmann 8 photographed the solar spec- trum at various altitudes from 116 meters to 4560 meters and concluded that the last trace of photographic action was independent of the altitude. They discovered two lines 291.67 mpi and 291.98 mi which had not been seen theretofore. 26 ULTRAVIOLET RADIATION The short-wave limits as determined by them for various altitudes at different places are shown in Table VIII. TABLE VIII Altitude Limit of solar spectrum 50 meters 291.26 m// 116 291.55 1620 291.36 3136 291.10 4560 291.21 They concluded that the limit on their spectrograms, where photographic action ceased, was constant but that the distribution of density near this limit varied some- what. Dember 9 investigated the limit of the solar spectrum by means of a quartz spectrophotometer and a photo- electric cell at an altitude of 4560 meters and found the limit at about 280 mji. He concluded that he obtained a shorter limiting wave-length because his photo-electric cell was more sensitive than the photographic plate. Wigand 10 used the same spectrograph which Miethe and Lehmann employed but reached the high altitude of 9000 meters. He also concluded that the limit of the solar spectrum was independent of altitude for he found the same limit at the earth's surface as at the high altitude which he attained where only about one-third of the atmosphere was above him. He found the last indication of photographic action at 289.7 m-p,. He attributed his lower value to the fact that he greatly reduced the fog on his plates by using a filter and thereby facilitated accuracy of measurement. This result further indicates a steep absorption curve due to the upper layers of the earth's atmosphere or to the sun's atmosphere. Fabry and Buisson " found the limit of the solar spec- trum slightly less than SOOmii and concluded that its abrupt ending was due to the absorption-band of ozone. SOLAR RADIATION 27 The maximum of the absorption-band of ozone is at 255 mji and it is quite marked between 200 mjj, and 290 mjj,. A depth of ozone of 25 pi at atmospheric pressure trans- mits only about one-half the radiation of a wave-length 255m\i. They recently 27 studied the solar spectrum be- tween 290 and 315mfi in great detail and found the inten- sity of the radiation at 290mpi to be only one-millionth as great as at SlSm^. Their work points conclusively to ozone as being responsible for the abrupt ending of the solar spectrum after passing through the atmosphere. The ozone in the atmosphere was determined to be equivalent to a thickness of 3 mm. at atmospheric pres- sure. They also showed that the location of this ozone was in the upper layers of the atmosphere, above 40 km. They conclude that the ozone is produced by solar radia- tion shorter than 200m|i, and that ozone is dissociated by radiation of longer wave-length, thus accounting for an equilibrium state. Kron 12 also studied the extinction of light in the terres- trial atmosphere in the ultraviolet region. Strutt 13 has discussed the transparency of the lower atmosphere for ultraviolet radiation and the relative poverty in ozone. He reviewed the work of Hartley 14 who suggested that the ultraviolet limit of the solar spectrum was due to ozone and also the work of Fowler and himself 15 which strengthened this view. He photographed the spectrum of the cadmium spark at a distance as great as 3600 feet and the spectrum of the quartz mercury arc at a distance as great as four miles. He used a quartz spectrograph and a small telescope containing cross-wires upon which the distant source could be focussed. The spectrum of the cadmium spark photographed through a horizontal distance of air 3600 feet in length extended to 231.3 m[i and apparently was not diminished at 255m[i where the absorption of ozone is a maximum. This indicates the absence of ozone or at least that it was not present in 28 ULTRAVIOLET RADIATION appreciable amount. The spectrum of the quartz mercury lamp obtained by a two-hour exposure, extended to 253.6 mji. The extent of this spectrum cannot be compared directly with that of the solar spectrum at sea-level because the equivalent thickness of air traversed in that case (zenith sun) would be more than five miles, whereas the distance Strutt used was four miles, but he compared it with data described by Cornu obtained from the peak of Teneriffe. A brief summary is given in Table IX for equivalent thicknesses of homogeneous atmosphere. TABLE IX Solar spectrum from near sea-level. . . . Solar spectrum from Teneriffe Mercury arc spectrum Thickness of air (feet) 29,000 17,900 20,100 Limit 294.8 292.2 263.6 It will be noted that the corrections being made for barometric pressure, temperature, etc., in order to obtain equivalent thicknesses of homogeneous atmosphere, Table IX, afford a comparison of the transparency of the upper and lower layers of the atmosphere provided the limit of the solar spectrum is due to an absorbing medium in the upper air. Strutt's conclusion is that the lower air is far more transparent than the upper air for ultraviolet rays shorter than SOOni^ if equal masses are considered, but this assumes the presence of an absorbing medium in the upper part of the earth's atmosphere which is still awaiting definite proof. However, the evidence is very strongly in favor of ozone as the medium which limits the solar spectrum. SOLAR RADIATION 29 According to Strutt's experiments there was no evi- dence that the limit of the mercury spectrum through four miles of lower air was necessarily at 253.6m|i. Longer exposures would probably reveal even shorter lines but this is not true of the solar spectrum. It ends so com- pletely that long exposures do not appear to extend it. Strutt made experiments on the absorption of radia- tion of wave-length 253.6mpi by ozone and concluded that 0.27ml! of pure ozone in four miles of air would suffice to produce the slight enfeeblement of this mercury line. Scattering of radiation by small particles acts in the same way as ozone to absorb ultraviolet radiation from a dis- tant source so that this complicates the quantitative estimations. Schuster 16 by using the Rayleigh 17 formula came to the conclusion that the selective absorption by the atmos- phere may be accounted for by the selective scattering due to molecules of air and the computed results agree favorably with the data obtained by Abbott at Mt. Wil- son. King 18 modified this formula and found excellent agreement between computed and observed results. Fowle 4 corrected King's formula and obtained excellent agreement as far as 37Qm\\, for observations made at mountain observatories where the atmosphere is fairly free from dust. However, this does not settle the matter of the ultraviolet radiation of shorter wave-lengths. Liveing and Dewar 19 determined the absorption of oxygen in a tube 165 cm. long under pressures of 85 and 140 atmospheres, obtaining the respective short-wave spectral limits, 266.4m[i and 270.4mji. This is the region in which ozone strongly absorbs, for its band is marked between 230mpi and 280mpi. They also used a tube of oxygen 18 meters long and a pressure of 90 atmospheres which provides about the same mass of the gas as is con- tained by the atmosphere and found the shortest wave- length transmitted was 366m\i. It has been suggested 30 ULTRAVIOLET RADIATION that perhaps their oxygen was not pure or that Beer's law does not hold for this region of the spectrum. The results obtained by various investigators upon the absorption by ozone indicate that much more ozone is re- quired to account for the limit of the solar spectrum than is present in the atmosphere near the earth. However, Pring 20 found a much greater concentration of ozone at an altitude of 3.5 kilometers than at the earth's surface. This indicates a possibility of sufficient ozone in the en- tire atmosphere to account for the limit of the solar spectrum. i ! j| ';! Abbott and Fowle 21 using a quartz prism, two mag- nalium mirrors, and a spectro-bolometer at an altitude of 14,502 feet above sea-level observed no appreciable energy of shorter wave-length than 290mji. In the high levels of the atmosphere there is very little moisture and the extreme ultraviolet radiation on passing through the dry oxygen may convert some or much of it into ozone. Where there is water-vapor present at ordinary temperatures the ozone would likely revert to oxygen. It appears certain that the presence of ozone in the upper atmosphere has been shown spectroscopically by several investigators. Water-vapor possesses an absorption-band in the ultra- violet region the maximum of which is in the neighbor- hood of 175m|,i. Air at atmosphere pressure and of a depth of 0.91 cm., transmits no appreciable amount of radiation shorter than 170mpi in wave-length. Under the same conditions oxygen transmits to about 185m|i. Columns of nitric and nitrous oxides about 20 cm. long transmits only about 12 per cent of radiation, 200mji in wave-length. Oxygen absorbs to about 186mpi at C. and when heated to 1800 C. the absorption-band extends to beyond 300mji. SOLAR RADIATION 31 Carbon-dioxide and nitrogen are practically transparent in the middle ultraviolet. This leaves oxygen as possibly responsible for the limit- ing of the solar spectrum but ozone appears to be the probable agency. Pure water is quite transparent to the near and mid- dle regions and is fairly opaque to infra-red. According to one observer 80 per cent of the solar energy is absorbed in the first meter of lake water and only about one per cent reaches a depth of four meters. Kowalski 22 obtained spectrograms of solar radiation re- flected by snow at an angle of 45 degrees. His exposures were made during four hours at midday at an altitude of 630 meters. The spectral limit was at about 295ml! thus showing that there was no appreciable selective absorp- tion. The reflection-factor of clean snow is more than 80 per cent. Ultraviolet radiation of short wave-lengths transforms oxygen into ozone and it is thought by some that the ozone formed in the upper regions of the atmosphere by solar radiation sinks toward the earth and oxidizes various impurities. An electroscope is quickly discharged by the influence of ultraviolet radiation of short wave-length but the effect is greatly reduced by interposing a quartz plate. This indicates the region of wave-lengths of the effective radiation. Various investigators have studied the photo-electric activities of solar radiation. Many years ago Bunsen and Roscoe conducted exten- sive photo-chemical researches which did much in supply- ing a foundation for photo-chemistry. They expressed the chemical effects of radiation in terms of chemical photo-units, each unit being determined by the chemical action upon a normal explosive mixture of hydrogen and chlorine contained in an isolation vessel of such small di- mensions that the variability of the extinction appearing 32 ULTRAVIOLET RADIATION in large vessels may be neglected when the explosive mix- ture is illuminated at a distance of one meter from a so- called normal flame. The normal flame burned carbonic oxide at a certain pressure at a platinum burner of certain dimensions. One "chemical light-unit" equalled 10000 of these photo-units. According to their formula, as pre- sented by Sebelien, 23 solar radiation reaching a horizontal area of the surface of the earth at an angle with the ver- tical will produce in one minute on each square unit of area a photo-chemical effect that may be expressed in " chemical light-units " by _ 0.4758P W = 318.3 (Cos0) 10" Cos *~ where P denotes atmospheric pressure, the constant 318.3 corresponds to photo-chemical intensity of solar radiation outside the earth's atmosphere, the constant 0.4758 de- notes the atmospheric extinction of direct solar radiation. 24 Sebelien 23 employing the formulae of Bunsen and Ros- coe, calculated the quantity of " actinic light which on the midsummer day falls upon a horizontal element of surface from sunrise to sunset " for various degrees of north lati- tude. His data for direct solar radiation, for diffused radiation from the sky, and for the sum of the two are presented in Table X using his terminology. Recently Karrer and Tyndall 25 have made an extensive investigation of the spectral transmission of the atmos- phere in the visible region. Their results indicate a gradual decrease in transparency toward the short-wave end of the spectrum. They made their measurements under various atmospheric conditions which may be characterized more or less approximately as follows: (1) clear sky and of low humidity; (2) overcast sky and of high humidity; (3) rainy. The average curve for the first general condition shows a gradual decrease in trans- parency with decrease in wave-length. That of the second SOLAR RADIATION 33 TABLE X Chemical " Light-units " per Unit Horizontal Area on Midsummer-day Chemical Light-units Degrees N. Lat. Direct Insolation Diffused Sky Radiation Total 60656 22060 82716 10 70891 23388 94479 20 77703 24539 102242 30 89060 25776 114835 40 79644 27059 106701 45 76178 27757 103935 50 72584 28521 101105 55 62704 28589 91293 60 62064 30484 92548 65 57089 32168 89257 70 50267 35012 85279 75 44587 37099 81686 80 40080 38612 78700 90 36211 39839 76048 condition shows two maxima, one at 580mpi and the other at 610mjx. The average curve for the rainy condition exhibits a maximum at 640mpi. Recently Bigelow 26 has published a treatise on solar radiation. References 1. M. Luckiesh, Light and Shade and Their Applica- tions, 1916, Chap. VII. 2. Smithsonian Meteorological Tables. 3. M. Luckiesh, Color and Its Applications, 1915 and 1921, 20. 4. Astrophys. Jour. 38, 1913, 392. 5. Astrophys. Jour. 17, 1903, 89. 6. Comp. Rend. 88, 1878, noi and 1285; 89, 1879, 808; 90, 1880, 940; in, 1890, 941. 34 ULTRAVIOLET RADIATION 7. Kayser's Handbuch III, 337. 8. Ber. Berlin Akad. 8, 1909, 268. 9. Abhand. Nat. Wiss. Gesell. Isis, Dresden, 2, 1912, i. 10. Phys. Zeit. 14, 1913, 1144. 11. Comp. Rend. 156, 1913, 782; Jour. d. Phys. 3, 1913, 196. 12. W. Schmidt, Mon. Weather Rev. (U.S.) Dec. 1914, 653. 13. Proc. Roy. Soc. 1918, 260. 14. J. Chem. Soc. 39, 1881, in. 15. Proc. Roy. Soc. A, 93, 1917, 577. 16. Theory of Optics, 1909, 329. 17. Collected Works, Vol. I, 87 and Vol. IV, 397. 18. Phil. Trans. Roy. Soc. Lond. A, 212, 1913, 375. 19. Kayser's Handbuch III, 361. 20. Proc. Roy. Soc. A, 90, 204. 21. Astrophys. Jour. 1911, 192. 22. Nature, March 30, 1911, 144. 23. Phil. Mag. 9, 1905, 352. 24. Pogg. Ann. 108, 257. 25. Bur. Stds. Sci. Pap. No. 389. 26. The Sun's Radiation. 27. Astrophys Jour. 54, 1921, 297. CHAPTER III TRANSPARENCY OF GASES In general most substances are increasingly opaque to ultraviolet radiation as the wave-length of the radiation decreases. This is one of the reasons for the difficulties encountered in the study of the extreme ultraviolet. In order to work with a degree of certainty with ultraviolet radiation, at least a small quartz spectrograph is indispen- sable. With such an instrument it is easy to determine the transmission characteristic of any substance as far as 200m^. However, some progress can be made without such an instrument if the general characteristics of re- flecting and transmitting media are known. At least the spectral limits are usually made fairly certain in this manner. Sources of ultraviolet and methods of measure- ment are discussed in other chapters. In this chapter there will be presented certain spectral characteristics of common media which are easily described or recognized. A vast amount of data is available to the author through the examination of material such as eye-protective glasses, dyes, and paints, but it does not appear worth while to in- clude data of this sort because of the uncertainty in the description of the substance due to the use of trade-names, etc. In the preceding chapter the spectral transmission characteristics of several gases were touched upon briefly in connection with the limit of the solar spectrum. For the sake of completeness they will be briefly discussed again, but the references will not be repeated. In general, the most extensive work in the extreme ultraviolet has been done by Schumann 4 and by Lyman. In the other regions there have been many investigators. 35 36 ULTRAVIOLET RADIATION Oxygen at atmospheric pressure possesses an absorp- tion-band, according to Lyman, extending from 127mjA to 176m|i. This band widens as the pressure increases. At a pressure of 40mm. it extends from 133m^ to 160m|ji and at 15 mm. it has diminished to a range from 135mpi to 150mjA. One investigator found an increase in the extent of the absorption-band with increase in temperature. At C the absorption-band ended at 186mpi but when the temperature was increased to 1800 C it extended into the near ultraviolet, that is, beyond SOOmp,. Liveing and Dewar employed oxygen in a tube 165 cm. long and found the short-wave limit of its transmission to be 266.4mpi at a pressure of 85 atmospheres and 270.4mjj, at a pressure of 140 atmospheres. They also found 1 that a depth of 18 meters of oxygen at a pressure of 90 atmospheres, did not transmit radiation shorter than 336mpi in wave-length* According to Kreusler 2 a column of oxygen 20.45 cm. long, at a pressure of 759 mm. and a temperature of 18.5 C absorbs 32.5 per cent at 186mji, 6.2 per cent at 193mpi and practically none at 200mpi. This indicates a sharp absorption band. Apparently it is oxygen that makes air opaque to radiations shorter in wave-length than 185mji. Lyman 3 has studied the absorption bands in detail. Kreusler's absorption-factors A (in per cent) including the absorption coefficients a in the usual equa- tion in which the transmission-factor equals e" ad where d is the thickness (in his case 20.45 cm.) are summarized herewith. m/z A a 186 32.5 0.02057 193 6.2 0.00336 Nitrogen is quite transparent to ultraviolet radiation, its slight absorption increasing gradually as the wave-length decreases. It is quite transparent even for radiation of wave-length 125mji. Schumann found it to be quite trans- TRANSPARENCY OF GASES 37 parent at 160mpi and Kreusler 2 determined its absorption to be only 2.2 per cent at 186mpi for a depth of 20.45 cm. at atmospheric pressure and room temperature. Lyman B used a column 9.14 mm. long at atmospheric pressure and observed only a slight absorption from ISOmfx to I25m\i. Hydrogen is quite transparent but it is difficult to study great depths of it in a pure state because of the impuri- ties arising from the container. Lyman by filling his " vacuum " grating spectroscope with hydrogen was able to obtain spectrograms for a depth of gas equal to about 200 cm. at pressures from 1 to 5 cm. An absorption-band was indicated at 170mji but this disappeared as the gas was renewed several times so it is likely that it was due to an impurity. A slight absorption was also observed between 130 and 133mjx. At atmospheric pressure the trans- parency of hydrogen ceased at about IGOmpi but Lyman was not certain of the purity of the gas after contact with the walls of the vessel. Later Lyman 6 reduced the pos- sibility of contamination to a minimum by using a depth of gas equal to 65 mm. He then observed that the hydro- gen transmitted radiation of wave-lengths almost to the short-wave limit of transparency of very transparent fluorite. It may be said that hydrogen is very transparent to ultraviolet radiation. Ozone was discussed at considerable length in Chapter II because of its possible relation to the abrupt ending of the solar spectrum. In the extreme ultraviolet the pres- ence of ozone does not appear to alter the absorption of oxygen. Apparently ozone is not particularly absorbing in the extreme ultraviolet. It possesses a powerful absorption-band with a maximum at 258mji, a minimum at 205m| > i, and extending markedly between 230m[i and 280m^. This band has been studied by several investi- gators and its presence and form is fairly well established. Meyer 7 using a photo-electric method obtained the values 38 ULTRAVIOLET RADIATION given in Table XI of the absorption coefficient, a, in the equation E = E 10~ ad . According to the curve in his original paper the absorption rapidly increases for radia- tion shorter than IQOmji in wave-length. TABLE XI Absorption coefficients of ozone m/x a m/i a 193 11.7 250 123.0 200 7.8 260 126.0 210 11.5 270 116.0 220 19.2 280 73.4 230 48.6 290 38.6 240 105.0 300 30.3 Ledanburg and Lehmann 8 evaporated liquid ozone and obtained a high percentage of gaseous ozone. In low concentrations they found that absorption extended to 316m|i but at higher concentrations bands appeared in the region of longer wave-lengths. The liquid ozone did not exhibit the ultraviolet bands. Helium exhibits about the same degree of transparency as hydrogen between 125m|i and 190mjA. According to Lyman no absorption in this region was observable. Argon is of the same order of transparency as helium and hydrogen. Carbon monoxide, according to Lyman, 5 exhibits eight narrow absorption-bands between 125m[i and 160mjx. It is more transparent to the extreme ultraviolet than car- bon dioxide. The bands are shorter for shorter wave- lengths and they decrease in width with a decrease in pres- sure of the gas. Its spectral transmission characteristic is quite complicated and apparently differs from any other gas which has been examined for transparency in the extreme ultraviolet. TRANSPARENCY OF GASES 39 Carbon dioxide exhibits some absorption-bands in the extreme ultraviolet. Shumann states that its trans- parency extends considerably further into the ultraviolet than that of oxygen. According to Kreusler 2 a column 20.45 cm. long at a pressure of 750 mm. and a tempera- ture of 15 C absorbs 13.6 per cent at 190m(i, 4 per cent at IQSmji, and 1.8 per cent at 200m|i. Kreusler's absorption- factors A (in per cent) and absorption coefficients, a, are as follows : m/z A a 186 13.6 0.00574 193 4.0 0.00213 200 1.8 0.00079 Water vapor appears to have an absorption-band with a maximum between 160mpi and 170mpi but apparently there are no very satisfactory results available. The transparency of air in the extreme ultraviolet is determined chiefly by its content of oxygen and in the middle ultraviolet and near the limit of the solar spectrum by its content of ozone. It is more transparent than oxygen at the same pressure. According to Lyman a depth of oxygen equal to 0.91 cm. is opaque to radiations shorter in wave-length than 176mjLi while the same depth of air transmits some radiation of wave-length 171mpi. The absorption of ultraviolet by air is one of the factors which makes it necessary to employ vacuum spectroscopes in studying the extreme ultraviolet. Kreusler 2 found that a column of air 20.45 cm. long at a pressure of 747 mm. and a temperature of 14 C absorbed 8.8 per cent at ISGm^ and no absorption was observable for radiation longer than 193mpi in wave-length. In other words air is quite transparent at 193mpi but almost opaque at ISSmji. Nitric and nitrous oxides were found by Kreusler 2 to absorb the extreme ultraviolet very strongly. His results 40 ULTRAVIOLET RADIATION for a depth of 20.45 cm. of nitric oxide at a pressure of 600 mm. and a temperature of 18 C are as presented in Table XII. Table XII Absorption by Nitric Oxide Wave-length mn Absorption-factor A Absorption-coefficient a 200 88.4 per cent 0.14932 210 76.3 0.09526 220 72.0 0.08424 230 54.6 0.05223 240 30.5 0.02406 250 4.7 0.00318 300 1.2 0.00083 Palmer 9 studied the " volume ionization " effect ob- served by Lenard 10 as produced by ultraviolet radiation of extremely short wave-length. It has been shown by Lyman 11 that when the secondary of a transformer is connected with additional capacity to the electrodes of a hydrogen-filled discharge containing traces of hydrocar- bons, the excitation of the tube gives rise to carbon bands extending as far as 170m^ into the ultraviolet region and to strong hydrogen lines from 125mji to 165m^. Palmer used such an arrangement with fluorite windows and a screen-cell containing oxygen. Lyman 5 has found that the absorption of radiation in the extreme ultraviolet is in the form of a band and that as the pressure increases the absorption spreads much more rapidly toward the less re- frangible side than in the other direction. For a column of gas 1 cm. thick at atmospheric pressure the band ex- tends from 126.8m^i to 177m[i and at a pressure of 0.02 atmosphere from 135mjj, to ISOm^i. Palmer admitted oxygen into the screen-cell at various pressures thus con- trolling the effective rays from the discharge tube. This TRANSPARENCY OF GASES 41 provides a " variable screen " for this particular part of the extreme ultraviolet region. Palmer found that the ionization of air, oxygen, and nitrogen is considerable but exceedingly small with hydro- gen. The power of ionization was found to increase greatly with decrease in the wave-length of radiation; at least this is true in the region of wave-lengths shorter) than 185mjx. The very large effect found with nitrogen may be due to a strong absorption of this gas for radia- tion between ISOmji and ISOmji. In experiments of this kind the Hallwach effect the ionization produced at an electrically charged surface when illuminated by ultraviolet radiation is confusing. Palmer eliminated this by covering the surfaces exposed to radiation with a film of soap solution. He also freed the gases from dust by admitting them through a long plug of cotton wool. Peskov 12 studied the spectral absorptions of chlorine and bromine and found that by varying a mixture of these two gases he could isolate regions of the spectrum as nar- row as 240mjx to 250mpi. These mixtures were found to obey Beer's law and therefore, after having obtained the requisite quantitative data, the mixture for filtering a certain spectral region could be calculated. According to Peskov, chlorine has a marked absorption between 300m|A and 400m^ as well as for the region of longer wave-lengths than 540mji. The absorption of bromine extends from 380mpi to 540mji therefore a mixture of the two gases pro- vides a screen for isolating the radiation of shorter wave- lengths than SOOmjj,. The very fine absorption-bands or lines exhibited by the vapors of some liquids were first noticed by Pauer. 13 Hartley 14 thoroughly studied those of benzene vapor and came to the conclusion that the ordinary broad bands were due to fusion of the fine bands. He recorded about 300 ultraviolet absorption lines. 42 ULTRAVIOLET RADIATION Baly 15 in applying the quantum theory concluded that the frequencies of the absorption-bands of a substance might be simple integral multiples of a fundamental fre- quency. According to this theory it would be possible by computation to predict unknown bands from known ones of different wave-lengths. He 16 found that in the case of benzene and p-xylene, the frequencies of the cen- ters of the groups of absorption, fluorescence, phosphores- cence, and cathodo-luminescence bands are represented by integral multiples of the frequency of an infra-red band. Baly calculated the wave-lengths of the absorption lines of benzene vapor and found that there should be about 600 lines between the limits where Hartley found 300 pro- vided the theory was correct. His computations agree well with the lines actually measured. He lays consider- able stress upon the constant frequency difference which he has observed for so many substances. 17 Stark 18 and his colleagues photographed the absorption spectra of a large number of hydrocarbons in the form of vapor as far as 185mpi and arrived at interesting con- clusions concerning the relation of linkings to absorption- bands. They examined the absorption spectra of hexane, cyclohexane, camphane, isobutylene, methyl butylene, hexylene, acetylene, diallyl, isobutylene, ethylene, methyl butadiene, dimethyl butadiene, methyl pentadiene, hexa- diene, bornylene, camphene, pinene, limonene, sylvester- ene, and a few other compounds. Strasser "investigated the ultraviolet absorption spectra of the vapors of several mono-substituted derivatives of benzene and found the absorption-bands to be similar for benzaldehyde, benzonitrile, benzyl alcohol, benzyl ethyl ether, and benzoic acid. Witte 20 determined the wave-lengths of the absorption- bands in the spectra of the vapors of benzene toluene, chlorobenzene, bromobenzene, aniline, phenol, and ani- sole, and estimated the relative intensities of the bands. TRANSPARENCY OF GASES 43 The spectra appear to be more or less similar. They pos- sess series of bands which exhibit a constant difference in frequency. Ribaud 21 determined the absorption coefficients of bro- mine vapor in the ultraviolet region. The vapor of carbon bisulphide exhibits an absorption- band in the ultraviolet which Paurer has resolved into lines. The vapor of ethyl benzene exhibits a number of absorp- tion-bands between 230 and 275m|i. Schulz 22 has discussed the work done on the ultraviolet absorption spectrum of benzene vapor and has presented the results of his own investigation. Using a concave grating and an iron arc he determined the positions of 75 bands. Between the members of a long series the mean difference in wave-length was found to be 9.2 lm^. The absorption of radiation by the vapors of selenium, tellurium, mercury, zinc, cadmium, phosphorus, arsenic, and bismuth has been studied by Dobbie and Fox. 23 Their investigation was confined chiefly to the visible region but extended into the near ultraviolet. A table of absorption bands is presented. Mercury vapor showed little absorp- tion at any temperature. Cadmium exhibited no general absorption but a few sharply defined bands occur in the ultraviolet such as: a very fine sharp band at 379.3m|i; a fine band at 326m^ which appeared first at about 600 C. and widened with increase in temperature ; a diffuse band at SlS.Gmji which appeared at 900 C; two sharp bands at about 365 and 370mji appearing at 1200 C; a band at SOe.lmjx appearing at 1000 C; and one at 338.2mpi which appeared at 1100 C. Zinc behaved in general like mer- cury but at 110 C four very sharp bands appeared at 369.9, 365.4, 338.4, and 328.4m[i respectively. No absorp- tion bands were observed for phosphorus, arsenic, and antimony although the general absorption increased with increase of temperature. 44 ULTRAVIOLET RADIATION The ultraviolet band of ammonia and its occurrence in the solar spectrum has been discussed at length by Fowler and Gregory. 24 Ribaud 25 has presented a discussion of the absorp- tion of radiation in different regions of the spectrum deal- ing chiefly with the broad continuous regions of absorp- tion shown by gases, liquids and solids. He refers to experiments which lead to the conclusion that for the same substance in different physical states or for the same chemical group in different compounds, the maxi- mum of the continuous region of absorption is more dis- placed toward the long wave-lengths the greater the value of the maximum absorption. Other experiments have shown that at a given temperature the damping in an ab- sorption band only depends on the position of this band in the spectrum. In other words if two bodies have an absorption band in the same region of the spectrum the dampings or the widths of their bands are the same. Ac- cording to Ribaud the width of an absorption band, which is solely a function of its position in the spectrum increases continuously on going from the ultraviolet towards the infra-red very nearly proportionally to the wave-length maximum. It is interesting to note that for all the ultra- violet and visible bands studied the observed widths of the absorption bands furnish a damping coefficient very approximately equal to the frequency. References 1. Phil. Mag. 26, 1888, 286. 2. Ann. d. Phys. 6, 1901, 418. 3. Astrophys. Jour. 38, 1913, 284. 4. Smithsonian Contribution, No. 1413, 29. 5. Astrophys. Jour. 27, 1908, 89. 6. Astrophys. Jour. 35, 1912, 344. 7. Ann. d. Phys. 12, 1903, 849. 8. Chem. Centr. 1906, 1727. TRANSPARENCY OF GASES 45 g. Phys. Rev. 32, 1911, i. 10. Ann. d. Phys. i, igoo, 486; 3, igoo, 2g8. 11. Astrophys. Jour. 23, igo6, 181. 12. J. Phys. Chem. 21, igiy, 386. 13. Wied. Ann. 61, i8g7, 363. 14. Phil. Trans. 208, igo8, 520. 15. Phil. Mag. 27, igi4, 632. 16. Phil. Mag. 2g, igis, 223. 17. Phil. Mag. 31, igi6, 425. 18. J. Chem. Soc. igi3, abs. 104, 363. ig. Z. Wiss. Photochem. 14, igi5, 281. 20. Z. Wiss. Photochem. 14, igis, 347. 21. Comp. Rend. 157, igis, 1065. 22. Zeit. Wiss. Phot. 20, ig2O, i. 23. Roy. Soc. Proc. g8, ig2O, 147. 24. Roy. Soc. Phil. Trans. 218, igig, 351. 25. Comp. Rend. 171, ig2o, 1134. CHAPTER IV TRANSPARENCY OF LIQUIDS Absorption-bands may be defined by their wave-length limits, the character of the edges, and the position of the center, but for any single solution, a curve showing the absorption-factors for various wave-lengths is most repre- sentative. Extinction-coefficients are also valuable be- cause it is then possible to compute absorption curves for other concentrations and thicknesses, provided Beer's law is valid as it is very generally. Hartley devised a method of representing by a single curve certain essential facts concerning the absorption spectrum of a liquid or of a substance in solution. The positions of the edges of the absorption-bands were determined for layers of different thicknesses and a curve was plotted which related wave- length (or frequency) and the thickness of the layer. These curves show the positions of the centers of the absorption-bands and of the edges for layers of any thick- ness. They show whether the bands are symmetrical or not and indicate the thickness of liquid necessary for ab- sorption to be evident. In order to show the absorption or transmission char- acteristics of a liquid or a substance dissolved in a solvent quite completely it is necessary to consider wave-lengths (or frequencies) of radiation, concentration (or depth) of the liquid, and absorption (or transmission) factors. This involves a figure of three dimensions. The author x has considered this figure graphically and has discussed vari- ous uses for spectral data. Hartley's method of plotting absorption data has been widely used although to make the curves more convenient 4 6 TRANSPARENCY OF LIQUIDS 47 in size the logarithms of thicknesses instead of the thick- nesses of the layers are usually plotted. The law relating thickness and absorption or transmis- sion, for radiation of a certain wave-length, may be ex- pressed thus: J = J A- cd or -f = A~ cd = -cd log A = log T Jo where J is the intensity of radiation entering the liquid or other substance, J is the intensity on leaving, A is the transmission coefficient, c is the concentration of the dis- solved substance, d is the depth of the layer, and T is the transmission-factor. The absorption-factor is found by subtracting the transmission-factor from unity unless there is loss by reflection or otherwise than by absorption. The author 2 has utilized this law in many practical ways to greatly reduce spectrophotometric and other spectral measurements. The absorption of ultraviolet radiation by dilute aqueous solutions of various salts was studied by Pidduck 3 by means of a photo-electric method. He employed a spark between zinc electrodes in a Leyden-jar discharge circuit. The radiation passed through a wire grating which formed the positive plate of a condenser, the nega- tive plate being of zinc connected to a pair of insulated quadrants of an electrometer. This is an application of the effect discovered by Hertz when he noted that the break- down voltage of a gap was less when the metal terminals were illuminated by ultraviolet radiation. Of course, there is much uncertainty as to the spectral character of the radiation although a fair estimate of the spectral lines and the limits of the spectrum may be made. There is a great decrease of the electrical action of ultraviolet radiation caused by transmission through or- dinary clear tap-water as compared with the effect after passage through the same thickness of distilled water 48 ULTRAVIOLET RADIATION Pidduck used not only distilled and tap-water but an arti- ficial tap-water consisting of small amounts of sodium chloride, magnesium sulphate, calcium sulphate, and cal- cium carbonate in distilled water. His control in each case was distilled water. The reduction in the electrical effect of ultraviolet radiation after passing through one of the solutions as compared to the effect after passing through distilled water depends upon the solution and the con- centration of the impurity. For example, a thickness of 15mm. of a solution of sodium whose concentration was 2.5 x 10 ~* normal, showed a reduction in electrical effect to 0.946 of the effect obtained when distilled water was used. When the concentration was increased to 200 x 10""* the electrical effect was reduced to 0.35 of the dis- tilled-water value. For 15mm. of ordinary clear tap- water the electrical effect as compared with that through dis- tilled water ranged from 0.114 to 0.173. Pidduck concluded that the absorption of ultraviolet radiation (the photo-electrically active rays) might be a sufficiently delicate test to distinguish between different kinds of distilled water but he was unable to detect any difference in various specimens of distilled water tested by him. However, the method may have some applica- tions. Kreusler 4 investigated the spectral transmission of distilled and ordinary tap-water. He found that water increased in absorption for the extreme ultraviolet on standing in a glass vessel and ascribed this increase to the dissolving of material from the vessel. For a thickness of 16.97mm. of distilled water he obtained the values given in Table XIII where A is the actual absorption-factor for this thickness of water and a is the absorption coefficient in the familiar equation. In the foregoing the absorption coefficient is a in the equation I = I e ad , where I is the intensity of the radiation before transmission, I the intensity after transmission, TRANSPARENCY OF LIQUIDS 49 TABLE XIII Absorption of ultraviolet radiation by water Wave-length m/x Absorption-factor A Absorption coefficient a 186 68.9 per cent 0.06884 193 24.5 0.01653 200 14.2 0.00899 210 9.8 0.00610 220 9.2 0.00567 230 5.6 0.00334 240 6.2 0.00316 260 4.2 0.00254 300 2.5 0.00151 and d is the depth or thickness of the layer of water in centimeters. Values of the absorption coefficient for radiation of longer wave-lengths obtained by Ewan, 5 Aschkinass, 6 and Nichols 7 are given in Table XIV. TABLE XIV Absorption of visible and near infra-red radiation by water Wave-length m/i Absorption coefficient a 415 0.00035 430 0.00023 450 0.0002 487 0.0001 500 0.0002 660 0.0003 600 0.0016 650 0.0025 779 0.272 865 0.296 945 0.538 50 ULTRAVIOLET RADIATION The general increase of the absorption coefficient with increase in wave-length is noticeable throughout the visi- ble spectrum. The sudden rise in the absorption coeffi- cient for the region of the near infra-red emphasizes the opacity of water for infra-red rays. Rubens 8 gives re- flection-factors, refractive-indices, and absorption coeffi- cients for water from Ipi to 18^. It is interesting to note the minimum at 487m[i. This accounts for the blue- green color of deep clear water. Hughes found 1 cm. of water to be opaque to radiation shorter than 220m^ in wave-length. Lyman 9 found that a prolonged exposure of a photographic plate to radiation passing through a fluorite cell containing a thickness of 0.5mm. of water showed a transparency for this depth of water as far into the extreme ultraviolet as 173mpi. Sodium, potassium, lithium, rubidium, caesium, barium, strontium, and magnesium react with liquid ammonia and produce excellent blue solutions. It is believed that the metal dissolves in the metal, forming a blue colloidal solu- tion but these colored solutions are not permanent. Absalom 10 dissolved these metals in ammonia and ob- tained blue solutions possessing valuable properties as ultra-violet filters. These solutions were quite fugitive but with dry ammonia and freshly scraped metal, Cottrell 11 obtained blue solutions lasting as long as several years. One general conclusion by Absalom was that trans- parency far into the ultraviolet region is much more com- monly met with in the case of color due to colloidal metals than it has been found to be in ordinary colored salts or aniline dyes.. Liquid ammonia is opaque to radiation shorter in wave-length than 240m|x. In all cases Absalom's blue solutions were opaque to radiation shorter than 244mji. A freshly prepared solution of col- loidal gold was opaque to radiation shorter than 249m|i in wave-length but after it had stood for about a day the limit of transmission was at 277mpi. TRANSPARENCY OF LIQUIDS 51 Argo and Gibson 12 studied the absorption of the blue ammonia solutions of sodium and magnesium for visible radiation and found that for the same intensity of color their absorption spectra are quite similar. They con- clude that this is strong evidence in favor of the assump- tion that the coloring principle is the same in both cases. The blue color of these and other metallic solutions ap- pears to be independent of the metal and the solvent used. Glycerine is nearly opaque to radiation beyond 230mji. Pfluger 13 found a thickness of 1 cm. to be opaque beyond 210mp, and absorption-factors as follows for radiation of other wave-lengths : Wave-length 227 257 275 293 330m^i Absorption 81 50 57 46 24 per cent. A depth of 1 cm. of chemically pure ethyl-alcohol was found by Pfluger to absorb the following percentages of radiation of different wave-lengths : Wave-length 203 206 214 219 227 240 280mpi Absorption 96 86 72 63 42 28 20 per cent Glatzel 14 has determined the absorption coefficients throughout the ultraviolet region for acetone, calcium nitrate, benzol, anthracene, and retene. Acetone is transparent to about 310mji and appears to have a maximum at 265m^ in the absorption-band which extends from 310 to 230mjx. Calcium nitrate is transparent to 250mpi but it absorbs more or less between 250 and 230m|ji. Apparently it is quite transparent to radiation longer than 310m^ in wave- length. Benzol is transparent to radiation of longer wave- lengths than 270m(i. Several sharp absorption-bands lie between 235 and 270mjj,. Anthracene exhibits several sharp absorption-bands be- tween 320 and 390mji. 52 ULTRAVIOLET RADIATION Retene begins to absorb at 300m^, the absorption in- creasing rapidly with decreasing wave-length. Canada balsam transmits ultraviolet radiation in a man- ner similar to ordinary glass. A thin film of the yellow- ish balsam may be said to be opaque to the middle and ex- treme regions ; that is, to radiation shorter in wave-length than about SSOmji. Glacial acetic acid is quite transparent to the near and middle regions. A layer 3 mm. thick transmits to the neighborhood of 200mji. Acetone, xylene, and turpentine differ somewhat in spec- tral transmission, but in general layers 3 mm. thick trans- mit only the near ultraviolet. Their spectral transmission characteristics are somewhat similar to those of ordinary glasses. Ether in a layer 3 mm. thick is quite transparent to about SOOmji and this layer is slightly transparent to the middle ultraviolet. Collodion in thin films is transparent to about 220mjx. Liquid ammonia is transparent to 240m[i. Liquid ethylene absorbs radiation shorter than 235mji in wave-length. Nitroso-dimethyl-aniline dissolved in water is fairly transparent between 280m|i and 400mjA. The best strength of this filter for isolating the near ultraviolet is one which just eliminates the blue and violet light. Inasmuch as this filter transmits other visible rays it is necessary to use such a filter as blue uviol (Jena) glass or a dye such as methyl violet contained preferably in a quartz cell or in a gelatine film on a quartz plate. If a red band is still trans- mitted this may be eliminated by a solution of copper sul- phate or other filter transparent to the near ultraviolet. Many combinations of aniline-dye solutions or dyed gelatine filters may be used for isolating portions of the near ultraviolet and visible regions. In general, gases and a few other media must be depended upon for isolating portions of the middle and extreme ultraviolet regions. TRANSPARENCY OF LIQUIDS 53 Yellow dyes are often used as photographic filters and for other purposes but they can not be assumed from their appearance to be opaque to the near ultraviolet radiation. For example, gelatine films on glass plates dyed with aurantia, tartrazine, fluorescein, aniline yellow, orange G and uranin, which appear approximately alike to the un- aided eye, differ markedly in spectral transmission in the near ultraviolet as shown by the author elsewhere. 15 These filters each consisted of 6.5 mm. of ordinary plate glass and a thin layer of dyed gelatine and each trans- mitted about 50 per cent of the (visible) light from a vacuum tungsten lamp. Three of them, aurantia, tartra- zine, and fluorescein, were opaque to the blue, violet, and ultraviolet; that is, to radiation shorter in wave-length than about 470mpi. The other three, orange G, uranin, and aniline yellow transmitted the radiation from the mercury arc in the near ultraviolet to a wave-length about 35 Om^. The orange G filter transmitted the blue and violet lines but the uranin and aniline yellow filters ab- sorbed these lines but transmitted the near ultraviolet to SSOmjj, fairly well. The radiation from the quartz mercury arc in the region of 366mpi can be isolated for photographic purposes when the ordinary plate is used by various combinations of solutions or gelatine filters of aniline dyes in which clear glass elements are used. A combination of aniline green and resorcine blue isolates 366m^ fairly well but acid green and ethyl violet appear to be better for the purpose. For photographic plates sensitive only to the blue, violet, and ultraviolet, or for other photo-chemical reactions in- volving only these radiations, dense filters of methyl violet and other deep purple dyes are quite satisfactory for isolating the region from 350mpi to 400mjx. Dense cobalt glass answers the same purpose. In these cases the visible red radiation is not effective and therefore need not be eliminated excepting in special cases. 54 ULTRAVIOLET RADIATION A solution of esculine is practically colorless (fluores- cing a pale blue) and is quite opaque to the ultraviolet region when contained in a glass cell or in a gelatine film on glass. In general, fluorescent solutions and solids are opaque to ultraviolet radiations although there are exceptions, es- pecially some of the feebly fluorescent solutions. Data pertaining to the spectral transmission-factors of many representative dyes have been published by the author elsewhere. 1 Spectrograms of radiation trans- mitted by many dyes and other substances, have been published by Uhler and Wood, 16 Mees, 17 and others. Wat- son's recent treatise 18 on the relation of color to chemical constitution contains a great deal of valuable material. Although most of the data pertains to visible and infra- red radiation, there are valuable glimpses of the ultra- violet region. For sharpening absorption-bands such absorption char- acteristics as that shown by neodymium ammonium ni- trate are very useful. For example, it is easy to obtain filters that will eliminate all radiation excepting the green and yellow lines of the mercury spectrum. By the use of a solution of the salt of neodymium or by employing a glass containing this element, the yellow lines of mercury may be completely eliminated leaving the green line. This is an excellent monochromatic radiation and it is of special interest because it is almost identical in wave-length to that of the most luminous radiation. Other mercury lines can be readily isolated as shown in Chapter VIII. A dilute solution of Auramine O in a quartz cell trans- mits as far into the ultraviolet as 250mpi but still absorbs the violet radiation between 400mpi and 450m[x. Dilute solutions of copper sulphate transmit the near ultraviolet and somewhat beyond SOOmji. A 1.5 per cent solution of cupric chloride in a quartz cell TRANSPARENCY OF LIQUIDS 55 5 cm. in depth transmits as far into the ultraviolet as 320mjx. Plotinkoff 19 investigated the absorption spectra of bro- mine and of cinnamic acid in benzene and determined the absorption coefficients of bromine in water, benzene, chloroform, and carbon tetrachloride. He used four lines of the mercury spectrum and concluded that bromine obeys Beer's law. He suggested that the bromine visible spectrum consists of two superposed absorption-bands. He also determined the absorption coefficients of aqueous solutions, erythrosine, acid-green, guina-green, potassium dichromate, and certain mixtures of dyes. He discussed the preparation of niters for isolating various spectral regions. Drossbach 20 obtained the ultraviolet absorption spectra of salts of rare earths, of alcohols, and of aromatic hydro- carbons. The spectra of the salts of erbium differ charac- teristically from those of didymium. The long-wave limit of absorption in the ultraviolet for various liquids are as follows: benzene, 290m^; toluene, 288mpi; xylene, SlOmpi; m-xylene, 307mji; mesitylene, 336mpi; propyl alcohol, ZQOmix; isobutyl alcohol, 335mji; amyl alcohol, 332mp,; allyl alcohol, SlOmjA. Methyl and ethyl alcohol are trans- parent throughout the near and middle ultraviolet regions. The results show that the presence of traces of the higher alcohols in methyl and ethyl alcohol can be detected by spectroscopy. Massol and Faucon 23 studied the absorption of certain alcohols for ultraviolet radiation. They found that the normal primary, secondary, and tertiary alcohols and three abnormal primary alcohols which they examined, ex- hibited a general progressive absorption for ultraviolet radiation. The tertiary alcohols were more transparent in the ultraviolet than the secondary and these, in turn, were more transparent than the primary. The three abnormal primary alcohols exhibited, in addition to the general ab- 56 ULTRAVIOLET RADIATION sorption characteristic of the normals, two absorption- bands in the region of 260mji and SlOmjx, respectively. According to these investigators, these bands are excep- tional, for they are not exhibited by the fundamental hy- drocarbon, by the other alcohols, or the corresponding alkyl haloids. The corresponding aldehydes possess only one broad band which lies between the two bands noted above. They also studied the absorption of radiation by the chlorides of ethane, acetylene, ethylene. None of these exhibited the characteristic band of chlorine. Tetra- chlorethylene absorbs more ultraviolet radiation than hexachlorethane or acetylene tetrachloride and this dif- ference appears to depend upon the saturation of the mol- ecule. Massol and Faucon 21 have investigated the transpar- ency of saturated aliphatic alcohols to ultraviolet radia- tion of the near and middle regions. They used as a source of radiation an arc between an iron electrode and one of brass coated with an alloy of tin, lead, and cad- mium. They obtained spectrograms over the range from 210 to SOOmpi. According to their results, ethyl and methyl alcohols are exceptionally transparent even to a depth of 10 cm. Propyl alcohol is of lesser transparency but the ab- sorption increases slowly with depth. Butyl, amyl, hexyl, heptyl, octyl, cetyl, and melissyl alcohols exhibit a rapidly increasing absorption as the depth of the layer increases in thickness up to about 1 cm. but as the depth is still in- creased beyond this, the transparency diminishes much more slowly. According to these investigators the trans- mission-factors decrease in general as the molecular weight increases. The limit of the spectrum for thick- nesses greater than one cm. is approximately SlOmpi for the primary alcohols with branched chains. They show two absorption-bands at about ZGOm^ and SlOmpi respec- tively for layers less than one cm. thick. In general, the order of transparency beginning with the least transparent TRANSPARENCY OF LIQUIDS 57 is as follows : primary alcohols with branched chains, nor- mal primary alcohols, normal secondary alcohols, tertiary alcohols. The same investigators studied the absorption spectra of the eosins 22 of certain confectionery colors 23 and of various dyes. The ultraviolet range was 220 to 405mn. They 67 have also shown that liquids and solutions which show absorption bands are the same which have been observed to exhibit magnetic birefraction. In some cases electric birefraction (Kerr-effect) goes together with absorption. Malachite-green, acid-green J, and Patent-blue exhibit a band in the visible, one near the boundary between the ultraviolet and visible, and one in the ultraviolet region. Water-blue 6B and acid-magenta each show a large band between 275 m|i and 320mjj,. Paris- violet and acid- violet 6B possess only the absorption-band in the visible region. The absorption spectra of the dyes derived from naph- thaleneazonaphthol were also studied. These included Bordeaux B, Crystal ponceau, Bordeaux S, Coccine, Fast red. The two dyes Ponceau RR and Scarlet R which are derived from xyleneazonaphthol were also included in the investigation. The spectra of these dyes are similar, con- sisting, in general, of a broad band from the yellow to the violet. The ultraviolet is strongly absorbed and in the greater concentrations all radiations shorter than the long- wave end of the visible spectrum are absorbed. Acid magenta is quite transparent to the violet and near ultraviolet of wave-lengths longer than 320mpi. For cer- tain concentrations a sharp absorption-band is present in the region between 270 and SOOmpi. Solutions of Orange I, Chrysoine, Naphthol yellow S, and Auramine O were examined in concentrations 1 : 10000. Auramine O is transparent more or less throughout the visible but has three bands in the ultraviolet located ap- proximately at 265, 310, and 350mji. Naphthol-yellow S 58 ULTRAVIOLET RADIATION transmits some green and the remainder of the visible of longer wave-lengths and it has an absorption-band in the ultraviolet at about SSSmpi. Orange I transmits radiation of wave-lengths longer than the sodium yellow lines, 589.3m^i, and has an absorption-band in the ultraviolet near 260m(i. The transmission characteristic of chrysoine for the visible spectrum lies between that of Orange I and Naphthol yellow S. According to Massol and Faucon a solution of fluores- cein exhibits three absorption-bands between 260 and 335mjx when of certain concentration and thickness; eosin, one band in the region of 335mjj,; and erythrosin and dichloroletraiodofluorescein (Rose bengal) each display an increasing absorption without bands. The concentra- tion of these solutions was 1 in 10000. Doubtless the fastness of dyes is related more or less to their absorption of ultraviolet radiation, but the relation is perhaps more simple and dependent for dyes of similar constitution. Certainly there are dyes of considerably greater permanency which absorb more ultraviolet energy than others which bleach rapidly. Meyer and Fischer * 4 have investigated the absorption in the ultraviolet region of solutions of fuchsone, benzaurin, Dobner's violet, salts of hydro xyphlenyl phthalide, dithio- fluorane, the alkali salts of quinizarin and purpurin, and several triphenylmethanes. Kruss 25 investigated the ultraviolet absorption spectra of the azo-compounds, the components of the azo dyes, and the derivatives of triphenylmethane. The solvents were ethyl alcohol, water, and in some cases sulphuric acid. The colorless bases and components of the dyes exhibit marked absorption-bands in the ultraviolet but these bands differ considerably for the various groups of dyes although for a given group the bands are quite similar, dif- fering chiefly, though slightly, in position. Apparently it is possible to determine the group to which a dye belongs TRANSPARENCY OF LIQUIDS 59 by means of a spectrogram of its absorption spectrum. Kruss strengthened the belief that the absorption of ultra- violet radiation increases with the number of double bonds in the molecule. He also reviewed previous work per- taining to absorption spectra of organic dyes. Dhere and Rhyncki 26 examined the spectra of the colorings, carotine and xanthophyll, and found that they are relatively trans- parent to ultraviolet radiation of longer wave-lengths than Bielecki and Henri 27 compared the absorption spectra of three fatty acids and the esters isomeric with them. The absorption of the number of groups of isomeric esters was determined in aqueous and alcoholic solutions for various radiations. They concluded that the absorption spectra of the various acids differ from those of the esters isomeric with them and that the difference is independent of the solvent. They also concluded that the absorption of ultraviolet radiation is controlled by the molecular com- plexity and increases with the complexity. A study of the four isomeric esters, butyl acetate, propyl proprionate, ethyl n-butyrate, and methyl valerate indicated that their absorption spectra vary considerably and are dependent on the molecular arrangement. They also studied the absorption by fatty acids and their esters in alcoholic solu- tions and by sodium formate and acetate in aqueous solu- tions and from the results they computed the molecular absorption coefficients. Their tabulated figures show that the absorption of ultraviolet radiation is almost the same for the esters as for the acids, the absorption of a com- pound of this type being determined by the acid group, the alcohol radical exerting small effect. Alcoholic solutions should show greater absorption than aqueous solutions. The acids in their order of increasing absorption are acetic, propionic, formic, butyric, and valeric. It is seen that the absorption increases with the addition of CH 2 to the mole- cule. The sodium salts which they studied exhibited less absorption than the acids themselves. 60 ULTRAVIOLET RADIATION Bielecki and Henri 28 have also extended their studies from monobasic fatty acids to polybasic saturated and un- saturated acids and their corresponding hydroxy acids. They conclude as a general result of their work that the effect of different chromophores in a molecule is not addi- tive but that the " absorption constant " is equal to the product of the " absorptive factors " corresponding to the chromophores and the " exaltation factors " which depend upon the relative position of the chromophores in the molecule. These values vary with the wave-length of radiation. The same investigators 29 determined the absolute values of the absorption of ultraviolet radiation for a num- ber of acids containing an ethylene and from a comparison of saturated acids determined the effect of such a linking. The ethylene linking in acids increases the absorption and the effect appears to increase as the linking approaches the carboxyl group. Bielecki and Henri 30 made a quantitative investigation of the ultraviolet absorption spectra of ethyl and methyl acetoacetates, ethyl ethylacetoacetate and diethylacetoace- tate, ethyl crotonate, mesityl oxide, pyruvic acid, ethyl puruvate, and ethyl laevulate. They have arrived at several conclusions from their studies of the influence of constitution on the absorption of ultraviolet radiation, es- pecially pertaining to the effect of various groups upon the increase in absorption and the displacement of the bands. The same investigators 31 studied the absorption of acetone and of aqueous and alcoholic solutions of acetone throughout the region between 215mpi and 370mjx and de- termined the absorption coefficients for radiations of vari- ous wave-lengths. A single absorption-band was found in each case, the maximum being at 270.6mpi for alcoholic solutions and at 264.8mj J i for aqueous solutions. Stark 32 investigated the conditions under which an in- flexion occurs in a spectral absorption curve and concluded TRANSPARENCY OF LIQUIDS 61 from the point of inflexion of the curve for acetone that it has a less intense absorption-band at about 330m|A. Bielecki and Henri 33 investigated the effect of the vari- ous groups containing nitrogen in aliphatic monamines, diamines, nitriles, carbylamines, oximes, and amides on the absorption of ultraviolet radiation. The absorption constant increases with decreasing wave-length as far as Ley and Fisher 34 studied the absorption spectra and fluorescence of the various imides. Succinimide exhibits absorption at about 400mpi and magnesium succinimide is still more transparent. The introduction of bromine into the molecule of maleinimide shifts the absorption- band toward longer wave-lengths. The presence of an amino group shifts the absorption-band toward longer wave-lengths and the addition of acid to solutions of amino-imides shifts the band toward shorter wave-lengths. Strobble 35 has shown that the red and the yellow modi- fications of fluorenone (diphenyleneketone) differ both in solid form and in solution. The absorption spectra of al- coholic solutions of these appear to be the same in the ultraviolet region. According to Bielecki and Henri 33 the aliphatic ketones and aldehydes possess absorption-bands at about 275mpi, the position depending upon the alkyl group. It shifts toward longer wave-lengths with increase in the number of CH 2 groups. Gelbke 36 studied the absorption spectra of several ke- tones in ethyl alcohol and water. The band of acetone whose maximum is at 268mji shifts toward longer wave- lengths when one or more hydrogen atoms is displaced by alkyl, phenol, halogen, nitroso, or carbonyl groups. Alkyl groups cause a displacement of 5 to lOmji and the effects of the other groups are greater. The shifting of the band is attended by a broadening of the band and an in- crease in absorption. 62 ULTRAVIOLET RADIATION Magini 37 has studied the absorption spectra of maleic and fumoric acids, asparagine, and the tartaric acids. He also investigated a number of aromatic compounds which exhibit strong absorption. These compounds in almost every case show distinct bands which were displaced to- ward longer wave-lengths when a hydroxyl group is re- placed by a carboxyl or an amino group. The introduction of a second carboxyl group into the chain appears to annul the increase in absorption and displacement of the bands resulting from the first. The isomerides rank in respect to increasing absorption in the order, meta, ortho, and para, respectively. Magini 38 has compared the absorption spectra of solu- tions of catechol, resorcinol, and quinol. They possess sharp absorption-bands in the region between 250 and 290mpi. Of the three, resorcinol exhibits the least absorp- tion and quinol the greatest. The absorption spectra of ortho and meta compounds possess the same maxima and minima but quinol differs from its isomerides in a peculiar absorption-band beginning at 250m[x. Henri and Wurmser 39 have reached an interesting con- clusion regarding the correctness of the Grotthus' photo- chemical law of absorption, which holds that " only the rays absorbed are effective in producing chemical change." They found that the maximum decomposition of acetone and ethyl acetate corresponds spectrally to the region of maximum absorption. The absorption curve of acetalde- hyde exhibits a maximum at 277.5m|i, then decreases to a minimum toward shorter wave-lengths and finally in- creases gradually in the region of still shorter wave- lengths. However, the decomposition is a maximum at 277.5mpi but it gradually decreases toward shorter wave- lengths, it is slight in the extreme region, and no percep- tible minimum of action is found between the maximum and the extreme region. Compel and Henri 40 have made a quantitative investi- TRANSPARENCY OF LIQUIDS 63 gation of the absorption of ultraviolet radiation by the three alkaloids, atropine, cocaine, and apoatropine, in al- coholic solutions. They determined the molecular con- stants of absorption for the maxima and the minima of the absorption spectrum. The spectrum of cocaine is quite different than that of the others. It exhibits a band at ZSlmjx which makes it possible to detect one part of co- caine in 200,000 of solution. One part of atropine can be detected in 2000 parts of solution and one part of apoatro- pine in 5000 parts of solution. Dhere 41 examined the near and middle ultraviolet spectra of various depths of solutions of some of the purine series. The spectrograms show that the extent of the ab- sorption spectra of the purines toward the longer wave- lengths increases with the amount of oxygen in the mole- cule. He 42 also examined the absorption spectrum of adreaniline in the ultraviolet region and found it to be similar to that for cathechol but its band is located at a slightly greater wave-length than the latter. Oxidation widens the band and displaces it toward the visible region. The absorption spectra of the purplish solutions of santonin and dihydrosantonin in alcoholic potassium hy- droxide, and of hydroxysantonins in alcoholic sodium ethoxide have been shown by Mayer 43 to be similar in the visible, the principal band being between 440 and 540mpi. However, the slight differences are accentuated in the ultraviolet region where the spectrum of santonin differs from the other two which resemble each other. Gibbs and Pratt 44 investigated the ultraviolet absorp- tion spectra of phenol, o-cresol, o-hydroxybenzyl alcohol, salicylic acid, and of methyl ether, of salicylic acid and methyl salicylate. They also noted the influence upon the absorption spectra of the addition of alkali. The ab- sorption spectra of benzyl alcohol, benzyl acetate, benzyl methyl ether, benzyl chloride, and methyl benzoate were also studied and it was found that the first four exhibit the same band, lying in the same region as that of benzene. 64 ULTRAVIOLET RADIATION The same investigators also photographed the absorp- tion spectra of o- and p-nitrophenols, and p-nitrosophenol, containing various amounts of sodium ethoxide. Purvis 45 studied the absorption spectra of alcoholic solu- tions of derivatives of benzoic acid to ascertain the influ- ence of substitution in the nucleus upon the absorption, and also the effect when the nucleus and acid radical are separated by saturated and unsaturated aliphatic groups as in phenylacetic, cinnamic, mandelic, and phenylpropi- onic acids. He studied the following derivatives substi- tuted in the nucleus : ortho-, meta-, and para-isomerides of toluic acid, chlorobenzoic acid, bromo benzoic acid, iodo- benzoic acid, and nitro-benzoic acid. The nitro group in benzoic acid is responsible for an extended though feeble band far out in the ultraviolet region. Purvis 45 also investigated the influence upon the ab- sorption spectra of halogen and nitrile derivatives of benzene and toluene of (1) the introduction in the ben- zene nucleus of the two dissimilar atoms of chlorine and bromine as in the o-, m-, and p-chlorobromo benzenes, as compared with the dichlor- and dibromo-benzene and with benzyl chloride; (2) the introduction of the nitrile group as in benzonitrile and the o-, m-, and p-toluonitriles, as compared wih phenylacetonitrile ; and (3) by the total re- placement of the hydrogen atoms in the nucleus, as in hex- achlorbenzene and hexamethylbenzene or by the addition of six atoms of chlorine as in hexachlorocyclohexane. Kober 46 has found that the absorption of aliphatic amino-acids in acid or alkaline solutions is only general in the ultraviolet. The aromatic amino-acids possess absorp- tion-bands which may aid in detecting them in pep tide chains. An excess of alkali increases the absorption and shifts the bands toward longer wave-lengths. The spectra of di- and tri-peptides exhibit no peculiar absorption. Baly and Hampson 47 have determined the absorption spectra of azobenzene, aminoazobenzene, dimethylamino- TRANSPARENCY OF LIQUIDS 65 azobenzene, and benzeneazophenyltrimethylammonium iodide. Baly 48 claims that the wave-lengths of the ultraviolet absorption-bands of phenol and aniline may be computed by using the constants which are characteristic of the in- fra-red spectra of benzene and water, and of benzene and ammonia, respectively. The agreement is quite satis- factory. Baly and Tryhorn 49 studied the absorption spectra of ethyl alcohol solutions of salicylaldehyde and aqueous solutions of pyridine of various concentrations. They found on increasing the quantity of solvent at first there results a shift of the absorption-band toward the visible region until a maximum wave-length is reached which de- pends upon the affinity between solvent and solute. Further dilution results in a shift of the band in the oppo- site direction until a minimum wave-length is reached. Cain 50 has shown that there is a similarity between the absorption spectra of p-benzoquinonediazide and of a-naphthalenediazonium chloride. Ley and Hegge 51 have examined the visible and ultra- violet absorption spectra of the cupric salts of the amino- acids, glycine, A- and B-alanines, piperidinacetic acid, anilonoacetic acid, aceturic acid, and a B-diaminopropionic acid. Hantzsch and Voigt 52 examined the ultraviolet absorp- tion spectra of the nitro compounds. They concluded that the true nitro-group exhibits feeble absorption at high concentrations, 0.1 to 0.01 normal; that the simple aci- group possesses a weak general absorption; and that the introduction of further unsaturated negative groups scarcely influences the absorption of true nitro-compounds but very greatly increases the absorption of aci-nitro-com- pounds and makes it quite selective in character. Garrett 53 employing a quartz spectrograph and a nickel spark examined the absorption spectra of solutions of sul- 66 ULTRAVIOLET RADIATION phurous acids, sodium sulphite, ammonium sulphite, sodium metabisulphite, potassium metabisulphate, rubi- dium hydrogen sulphite, potassium hydrogen sulphite, acetone sodium hydrogen sulphite, and potassium sodium sulphite. According to Stark and Levey 54 the absorption spec- trum of benzene displays a group of bands between 230 and 270mjx and a group of stronger bands between 190 and 210m|j,. Napthalene displays similar groups, the more in- tense one lying between 190 and 220m[i and the less in- tense one between 230 and SlOmpi. It will be noted that those of napthalene are of greater wave-lengths than those of benzene. Wiemer 55 examined the absorption spectrum of ethyl benzene between 230 and 275mji for the vapor and for solutions in ethyl alcohol. In both cases there were series of bands diminishing with increasing wave-length. An increase in temperature broadened the bands toward longer wave-lengths without affecting the short-wave edges. The absorption spectrum of toluene appeared similar to that of theyl benzene but the bands were slightly displaced toward the shorter wave-lengths. Pfliiger 56 determined the absorption-factors of several ethereal oils and synthetic organic substances for various lines of the mercury spectrum. The substances were placed in a quartz cell and the energy was measured by means of a thermopile. Bielecki and Henri 57 determined the absorption spectra of aliphatic alcohols, acids, ketones, aldehydes, and esters. The alcohols exhibited a progressively increasing absorp- tion from 300 to 214mjA which was augmented by the addition of CH 2 groups in the molecule. Oxalic acid was found to be 30,000 times and the monobasic acids about 2000 times, more absorbing than methyl alcohol. Rosanoff 58 examined the ultraviolet absorption spectra of hydrogen peroxide and concluded that it was quite TRANSPARENCY OF LIQUIDS 67 probable that the absorption of radiation of short wave- lengths by radioactive substances is partially due to the hydrogen peroxide formed by the emanation. Shaefer, Higgemann, and Kohler 59 studied the absorp- tion spectra of sulphurous acid and its salts and esters, sulphur dioxide, chlorous acid, and the chlorites, nitric acid, nitrates, and alkyl nitrates, and hypochlorous acids and its salts and esters. They employed a large range of concentrations. The absorption of the halogens and their acids for ultra- violet radiation has been studied by Cohen and Stuck- hardt 60 using quartz, uviol glass, and ordinary glass. Vallet 61 in studying the effect of ultraviolet radiation from a quartz mercury arc on bacillus coli communis in various solutions found that ethyl alcohol, glycerol, and certain saline solutions are quite transparent to the effec- tive radiation. However, peptone, albumen, and oil were among those liquids which were found to be quite opaque to the effective radiation which is, in general, of shorter wave-length than SOOm^. The transparency of various coloring matters in differ- ent concentrations has been studied by Miethe and Sten- ger 62 by means of a quartz spectrograph. The region of maximum transparency of tartrazine solutions increases from 300-308mpi in the strongest solution to 280-391m|i in the weakest solution. The concentrations were varied from 1:1000 to 1:20000. For similar concentrations of filter yellow the corresponding regions are 296-308mpi and 270-500mjx respectively, and for Martius yellow 321- 330mji and 296-374mpu For a concentration of 1 :9000 of nitrosodimethylaniline the maximum transparency is at 299-3 65 mpi and increases only slightly with dilution. For a concentration of 1:1000 of eosin the maximum trans- parency is 368-390m^i and for 1 :10000 this region increases to 271-470mpi. With increasing dilution fluorescein in- creases in transparency down to 26Qm\i. They also in- 68 ULTRAVIOLET RADIATION vestigated the well-known transmission-band of a silvered quartz mirror and found for exposures increasing from 5 to 640 seconds the region of transparency increased from 308-330mji to 302-388m[i. They concluded that the mir- ror is decidedly less suitable as a filter than the dyestuff filters. Stumpf 63 in a search for filters which transmit ultra- violet radiation examined a number of yellow dyes and found that flavazin L gave the best result. It is fairly transparent between 290 and 320m|i. The absorption of ultraviolet radiation by unsaturated compounds recently has been further treated by Ley. 64 This work was especially concerned with the anomalous hypochromous effects of the alkyles in the case of deriva- tives of styrene, stilbene, and cinnamic acid. With the introduction of alkyles into alpha-compounds one absorp- tion band of styrene vanishes and the other is displaced toward shorter wave-lengths contrary to expectation. In beta-compounds the displacement is in the opposite direc- tion. The work was done by the method of Hartley-Baly and iron arcs and alcoholic solutions were employed. The effect on the ultraviolet absorption of acetone and its homologues in a solvent has been investigated by Rice. 65 The results show that all the aliphatic ketones ex- cept acetone and methyl ethyl ketone follow Beer's law, the molecular extinction being independent of the con- centration and of the solvent. With these exceptions there are deviations when ionizing solvents are used owing to partial disruption of the associated molecules. Kundt's rule, according to which the absorption band is displaced toward the red as the solvent increases in refractivity, holds good as a rough generalization. When a pure sub- stance is dissolved in a solvent of ionizing type, the ab- sorption center moves toward the ultraviolet, whereas if the substance is dissolved in a solvent of neutral, non- ionizing character the absorption center either remains un- TRANSPARENCY OF LIQUIDS 69 affected or is displaced toward the red end of the spec- trum. The conclusion is drawn that this is probably a general rule valid for absorbing substances of all classes. Orndorff, Gibbs and Scott 66 found the transmission of one cm. of absolute ethyl alcohol in the region from 240 to 370mpi to be less after boiling than before being heated. It is possible that the change in transmission is not due to any real change in absorption but to scattering produced by colloidal particles of the material of the flask or to oxidation of the alcohol. References i. Color and Its Applications, 1921 ; J. Frank. Inst. 184, * 73 and 227. a. Trans. I. E. S., 9, 1914, 839. 3. Phil. Mag., 17, 1909, 710. 4. Ann. D. Phys., 6, 1901, 418. 5. Proc. Roy. Soc., 57, 1894. 6. Wied. Ann., 55, 1895. 7. Phys. Rev., i, 1893, i. 8. Ladenburg, Verh. d. Phys. Ges., 1909, 19. 9. Nature, 8, July, 1910. 10. Phil. Mag., 33, 1917, 450. 11. J. Phys. Chem., 1914. 12. Phys. Rev., 7, 1916, 33. 13. Phys. Zeit., 5, 215. 14. Phys. Zeit., i, 1900, 173 and 285. 15. Trans. I. E. S., 9, 1914, 472. 16. Atlas of Absorption Spectra. 17. Atlas of Absorption Spectra. 18. Color in Relation to Chemical Constitution, 1918. 19. Z. Phys. Chem., 79, 1912, 357. 20. Ber., 35, 1902, 1486. 21. Bui. Soc. Chem., n, 1912, 931. 22. Bui. Soc. Chem., 13, 1913, 217. 23. Comp. Rend., 157, 1913, 206, 386, 513, 700. 24. Ber. 46, 1913, 46, 70, 92. 70 ULTRAVIOLET RADIATION 25. Z. Phys. Chem. 51, 1905, 257. 26. Comp. Rend. 157, 1913, 501. 27. Comp. Rend. 155, 1912, 1617; 156, 1913, 550. 28. Ber. 46, 1913, 2596. 29. Comp. Rend. 157, 1913, 372. 30. Comp. Rend. 158, 1914, 866 and 567. 31. Comp. Rend., 156, 1913, 884. 32. Phys. Zeit. 14, 1913, 845. 33. Comp. Rend. 156, 1913, 1860. 34. Ber. 46, 1913, 327 and 3627. 35. Ber. 44, 1911, 1481. 36. J. Chem. Soc. 1913, abs. 104, 87. 37. J. Chem. Soc., 1904, abs. 86, 107. 38. J. Chem. Soc. 1903, abs. 84, 706. 39. Comp. Rend. 156, 1913, 230. 40. Comp. Rend. 156, 1913, 1541. 41. Comp. Rend. 141, 1905, 719. 42. Bui. Soc. Chem. i, 1907, 834. 43. Atti. Accad. Lincei, 1914, 23, i, 422. 44. J. Chem. Soc. 1915, abs. 108, 500. 45- J- Chem. Soc. 107, 1915, 966 and 496. 46. J. Biol. Chem. 22, 1915, 433. 47. J. Chem. Soc. 107, 1915, 248. 48. Phil. Mag. 30, 1915, 510. 49 J- Chem. Soc. 107, 1915, 1121. 50. Ber. 46, 1913, 101. 51. Ber. 48, 1915, 70. 52. Z. Phys. Chem. 79, 1912, 592. 53. J. Chem. Soc. 107, 1915, 1324. 54. J. Chem. Soc. 1913, abs. 104, 366. 55. Z. Wiss. Photochem. n, 1913, 33. 56. Phys. Zeit. 10, 1909, 406. 57. Comp. Rend. 155, 1912, 456. 58. J. Chem. Soc. 1912, abs. 102, 875. 59. Z. Elecktrochem. 21, 1915, 81. 60. Z. Phys. Chem. 91, 1916, 722. 61. Comp. Rend. 150, 1910, 632. 62. Zeits. Wiss. Phot. 19, 1919, 57. TRANSPARENCY OF LIQUIDS 71 63. Zeit. Wiss. Phot. 20, 1921, 183. 64. Zeits. Wiss. Photochem. 18, 1919, 177. 65. Am. Chem. Soc. J. 42, 1920, 727. 66. Phys. Rev. 19, 1922, 393. 67. Soc. Chemie Bui. 25 and 26, 1919, 585. CHAPTER V TRANSPARENCY OF SOLIDS The transparency limits and spectral transmission char- acteristics of various solids are important in studies and applications of ultraviolet radiation. In case a quartz or reflection-grating spectograph is unavailable, data per- taining to the media employed are useful in indicating the spectral limits of the radiation involved. Furthermore, various media provide a means of eliminating different spectral regions and thus provide a means of systematic investigation. Owing to the extensive use of glasses these are discussed separately in Chapter VI. The present chapter deals with other solids. Although the transparency of quartz extends suffi- ciently into the ultraviolet region to be suitable for opti- cal systems in the study of the near and middle regions of the ultraviolet, it was clear colorless fluorite that aided Schumann in invading the extreme region. Lyman x found clear colorless fluorite, 1 to 2 mm. thick, to transmit radia- tion of wave-lengths as short as 125mpi. Specimens of fluorite vary considerably in their spectral range of trans- parency. Hughes 2 examined thirty specimens and found only five which transmitted further into the ultraviolet than crystalline quartz. Two specimens of clear colorless fluorite showed limits at 133mji and 170m|j, respectively. This emphasizes the necessity of testing each specimen regardless of its appearance. Lyman examined specimens, 1 to 2 mm. thick, cut from pink, green, purple, and yellow fluorite and found their transparencies to vary considerably. Of fifty-seven speci- mens, forty- two showed limits of transparency to be at 72 Quartz Mercury Arc Giuartz Plate Photographic Bulb Daylight Bulb Clear Bulb Pyrex G/ass Hard Glass So-Ft Glass Plate Glass Window Glass Gelatine Celluloid Mica Quartz Mercury Arc CM Plate II. The Transmission of various media for ultraviolet radiation from the iron arc as obtained by a quartz prism spectrograph. Spectra of the quartz mercury arc shown for comparison. TRANSPARENCY OF SOLIDS 73 longer wave-lengths than 170mpi; ten were about equal in transparency to quartz 1 mm. thick; and five were nearly as good as his best specimen noted above. He concluded that deep coloration was a fair indication of limited transparency but certain greenish specimens were exceptions, they being as nearly transparent as the speci- men which transmitted to 125mpi. By heating, the color of fluorite can be made to disappear and there is usually a consequent gain in transparency. At the present time the best specimens of fluorite are the most transparent known solids for the extreme ultraviolet. Quartz ranks next to fluorite as a solid transparent to ultraviolet radiation. There is considerable difference be- tween crystalline and fused quartz, the former usually being transparent further into the ultraviolet than the fused silica. The crystals vary considerably in trans- parency, the best specimens being transparent to radiation as short as 145mji in wave-length. Lyman 3 found a piece of crystalline quartz 0.2 mm. thick to transmit radiation of wave-lengths as short as 145m|i. For a thickness of 2 mm. the transparency had shrunk to ISOmji and a speci- men 20 mm. thick was opaque to radiation shorter than 160mpi in wave-length and was almost opaque to radia- tion between 160m|i and 200m^ in wave-length. It may be that 200mjx is the limit of transparency of quartz spectro- scopes owing to the great depth of quartz through which the radiation must travel. Pfliiger 4 found variations in the transparency of quartz crystals and a generally lower transparency when the path of radiation was parallel to the axis than when perpendicular. For a specimen 1 cm. thick he obtained the following values of absorption : Wave-length 186 203 214 222 m/* Absorption 33 16 8 6 per cent Fused quartz varies widely in transparency as does the natural crystal. Hughes 2 found fused quartz 0.3 to 0.4 mm. 74 ULTRAVIOLET RADIATION thick, to transmit 24 per cent of the radiation of wave- length 185mpi; 36 per cent of 197mpi; and 40 per cent of 200mpi. This rapid decrease in transparency indicates the reason for 185mji being about the short-wave limit of ra- diation emitted by the mercury arc enclosed by a fused silica tube. The mercury line 184.9mpi can be obtained from a quartz mercury arc whose tube consists of fused silica nearly 1 mm. thick. Pfliiger * examined specimens of fused quartz and for example, found a specimen 2.8 mm. thick to be opaque to radiation shorter than 200mji in wave-length. Lyman 8 in search of material which would be trans- parent enough for extreme ultraviolet radiation in order that it might be used for the construction of windows for vacuum tubes, etc., examined many substances. He found borax, adularia, calcite, chrysoberyl, sanidin, arragonite, apophyllite, silver chloride (horn silver), diamond, and kunzite quite opaque to radiation of shorter wave-length than 200mji and some of them opaque to even longer wave- lengths. Gypsum, 1 mm. thick and bounded by cleavage surfaces transmitted radiation of wave-lengths slightly beyond Colemanite was opaque to radiation shorter than 175m[x in wave-length. Sugar, 1 mm. thick cut from " rock candy " was less transparent than colemanite. Barite, 1 to 2 mm. thick was opaque to radiation of wave-lengths shorter than 175m|x. Alum, 1 mm. thick, appeared to be more transparent than barite but its spectrum also ended at about 175m^. Celestite was transparent down to about 170m[i. Lyman found a specimen of topaz from Ceylon to pos- sess great transparency for the extreme ultraviolet, its transparency extending to about 157mji for a specimen 1.5 mm. thick. He found topaz from Japan, Siberia and TRANSPARENCY OF SOLIDS 75 Utah much less transparent than the specimen from Cey- lon but this may be due to a peculiarity of the individual specimen. Pfluger 4 found calcspar 6.1 mm. thick to absorb as follows : Wave-length ... .214 231 240 258 280 m/i Absorption 97 69 44 26 15 per cent Rock-salt possesses much higher dispersion than quartz or fluorite, therefore Pfluger recommended rock-salt prisms protected from disintegration by air by thin quartz plates cemented with glycerine. However, Lyman found rock-salt to be less transparent to the extreme ultraviolet radiation than quartz and the absorption to increase rapidly beyond 180m(x. A specimen 2 mm. thick became opaque at lyymji. Pfluger obtained the following values of absorption-factor for a thickness of 5.65 mm. of rock salt: Wave-length 186 210 231 280 m/* Absorption 30 23 14 4.5 per cent Absalom 5 determined the short-wave limit of transmis- sion for various gems and minerals. He employed an arc between copper poles and a small quartz spectrograph. His results which apply, of course, to his particular speci- mens, are presented in Table XV. The wave-length values represent the wave-length in each case at which complete absorption begins; that is, they represent the short-wave limits of transparency. The natural-blue rock-salt was said to be mined at the boundary of the salt with some of the potash minerals. A piece of ordinary colorless rock- salt was colored a deep blue by cathode rays. It appeared to have about the same transparency for ultraviolet radia- tion as the natural-blue specimen. The color induced by cathode rays was present only at the surface. The violet specimen of Chili saltpetre, occurring naturally, can be decolorized by heat. The author 6 has shown this to be 76 ULTRAVIOLET RADIATION true of the purplish tinge brought out in clear glass (con- taining small amounts of manganese) by the influence of ultraviolet radiation. After decolorization by heat the violet tinge can be restored to the saltpetre by exposure to cathode rays. Likewise the purplish tint can be brought out in clear glass, containing manganese in small quanti- ties, by X-rays. 6 Absalom concluded that there is a strong likelihood of the coloring in many of the substances which he examined being due to colloidal metals. TABLE XV Short-wave Limits of Transparency of Various Substances m/j 225 225 225 225 225 Natural blue rock-salt Beyond Rock-salt colored blue by cathode rays Beyond Sylvite, white (native potassium chloride) Beyond Sylvite, colored blue by cathode rays Beyond Fluorspar, colored violet by cathode rays Beyond Chili saltpetre, white 351 Chili saltpetre, violet 325 Diamond, yellow 320 Diamond, blue 315 Kunzite 305 Garnet, red 402 Zircon (hyacinth) red-brown 262 Zircon decolorized by heat 244 Zircon, green 402 Zircon, yellow 402 Topaz, pale yellow 262 Topaz, dark yellow 229 Topaz, pale pink-brown 262 Topaz, blue 296 Emerald 320 Ruby 300 Tourmaline green 351 Tourmaline green-yellow 300 Tourmaline pink 306 Spinel blue 402 Spinel purple 325 Spinel pink 300 Kyanite blue 320 Beryl blue 327 Cordierite blue-purple 325 Cairngorm 325 TRANSPARENCY OF SOLIDS 77 Gelatine in thin layers is quite transparent to the near and middle ultraviolet but it is quite opaque to radiation of shorter wave-lengths than 200m[i. For this reason non-gelatine photographic plates must be used for photo- graphic investigations in the extreme ultraviolet region. Gelatine filters may be made by incorporating the dye in a solution of 6 grams of gelatine to 100 grams of water heated to about 50 C. This is flowed upon a level plate of quartz, glass, or other material, and allowed to dry. A fixed photographic plate may be dyed successfully by al- lowing it to soak in a dye-solution. In fact, this is usually the best way for those lacking in experience with gelatine solutions, provided the filter is to be used only in the region where glass is transparent. A film of silver may be deposited of such thickness that it will transmit a narrow band between 310 and SSOmpi and still be fairly opaque to the remainder of the ultra- violet region. Miethe and Stenger 9 found for exposures increasing from 5 to 640 seconds the region of trans- parency increased from 308-330mpi to 302-388m|i. Dense cobalt-blue glass isolates the near ultraviolet for many photo-chemical processes such as ordinary photog- raphy. There is usually a deep red band transmitted but this can be absorbed by a solution of copper sulphate. The latter also absorbs much of the infra-red radiation. Cobalt oxide is one of the few coloring materials which does not decrease the transparency to ultraviolet radiation. In fact the author 7 has shown positive evidence of its ex- tending the limit of transparency of glass farther into the ultraviolet. Thin sheets of " celluloid " and photographic film are transparent to the near ultraviolet but are fairly opaque beyond 300m|x. Mica varies considerably in transparency but in general it strongly absorbs ultraviolet radiation of short wave- lengths. Specimens examined are similar to ordinary 78 ULTRAVIOLET RADIATION glass in spectral transmission, being quite opaque beyond SOOmjj,. Specimens of mica can be obtained in various colors and these are in general more opaque to ultraviolet radiation than the colorless specimens. Doelter 8 has discussed the coloring matter of various minerals including precious stones. The cornea of the human eye is opaque to radiation shorter than 295mpi in wave-length; slightly absorbing for radiation between 295 and 315m^ in wave-length; and quite transparent to radiation of longer wave-lengths be- tween 315 and 750m[i. The lens of the adult human eye is opaque to radiation shorter than 376mpi in wave-length and usually to all ultraviolet radiations. Apparently the lens of the eye of a child transmits slightly between 315 and 330mjA. References 1. Spectroscopy of the Extreme Ultraviolet, 1914, 58, 2. Photo-electricity, 1914, 137. 3. Astrophys. Jour. 25, 1907, 47. 4. Phys. Zeit. 5, 1904, 215. 5. Phil. Mag. 33, 1917, 450. 6. Luckiesh, Gen. Elec. Rev. 20, 1917, 671. 7. J. Frank. Inst. 186, 1918, in. 8. Monatsh. 30, 1909, 179. 9. Zeits. Wiss. Phot. 19, 1919, 57. CHAPTER VI TRANSPARENCY OF GLASSES The many kinds of glass are so widely used that it ap- pears of interest to discuss them in a separate chapter. In general, it may be said that what is termed " clear " glass is opaque to the middle and extreme ultraviolet. Most of them are quite transparent to the near ultraviolet to the neighborhood of about SSOm^ but from this point their transparency rapidly falls off to practically zero at SOOmji. A spectrogram of skylight or of sunlight taken with ordinary window glass intervening will show a short- ening at the short-wave end as compared to the spectro- gram taken through the open window. As has been seen in Chapter II, the short-wave limit of solar radiation is in the neighborhood of 295mji for a point near sea-level. Ordinary window glass cannot be considered to transmit radiation of this wave-length; however, there are plenty of glasses of appreciable thicknesses whose transparency to visible radiation persists undiminished to the region of 350mji. The spectral transmission curves of " clear " glass show rather sudden changes; that is, when appreciable absorption begins it increases rapidly with wave-length. Naturally, the manufacturers of glass are reluctant to liberate data pertaining to the chemical constitution of their products. The transmission characteristics can be determined readily by means of a quartz spectrograph and other apparatus, but there usually remains an indefinite- ness due to the use of trade names and other terms which do not even approximately specify the contents of the glass. The transparency of so-called flint glasses ordinarily does not extend quite as far into the ultraviolet as that 79 80 ULTRAVIOLET RADIATION of crown glasses. In other words, glasses of high re- fractive-index are likely to be less transparent in the region of SlOmji as those of lower refractive-index. In Table XVI are found the short-wave limits of trans- parency (or of absorption) of certain representative clear colorless glasses 2 mm. in thickness. TABLE XVI Refractive-index Absorption limit Common glass . 295mju Light crown 1.61 296 Extra light flint 1.64 298 Medium crown 1.62 300 Light flint 1.57 306 Tvledium flint 1.62 315 Extra dense flint 1.69 335 Schott's heavy flint 340 The variation in the short-wave limit of transparency of glasses can be utilized to great advantage in controlling the spectral range of radiation in the near ultraviolet re- gion. By means of a series of glasses such as those repre- sented in Table XVI it is possible to explore the proper- ties of the near ultraviolet even when the total radiation is used, that is, when no spectroscope is employed. For example, if a high intensity of the near ultraviolet radia- tion is desired, a quartz mercury arc, iron arc, or other source may be used with a screen of light crown glass. After noting the effect on the phenomenon under investi- gation, a thicker specimen or a medium crown glass may be used. Thus gradually the near ultraviolet spectrum may be shortened at the short-wave end with obvious possibilities in systematic investigation. In a manner simi- lar to this it is possible to determine the amount of total ultraviolet energy radiated by various sources. One of TRANSPARENCY OF GLASSES 81 the essentials is a screen which cuts off only the ultraviolet radiation. By comparing the energy with and without the screen the difference will yield the desired values. For this latter purpose Euphos glass has been used, and as will be seen in Table XVIII, other glasses of the proper thickness answer the purpose. The practical limit of transparency depends upon the thickness as well as upon the particular kind of glass, so that a given specimen may be ground to the thickness which suits the purpose best. The loss by reflection of perpendicularly incident radia- tion at a polished surface of glass can be computed by means of Fresnel's formula provided the refractive-index n is known. This loss is about four per cent for each surface of light crown glass and nearly seven per cent in the case of heavy flint glass. This statement pertains only to the region of normal dis- persion and to yellow light. The amount of reflected radiation being dependent upon the refractive-index and the latter varying with the wave-length of radiation, it is obvious that the amount of reflected radiation depends upon the wave-length. For determining the loss by re- flection for smaller angles of incidence than 90 degrees the more general formula involving the angle of incidence is used. This can be found in treatises on optics. The percentage of reflected light increases slowly with increas- ing angle of incidence up to 60 degrees. It then increases rapidly, becoming, of course, very great at angles of in- cidence near 90 degrees. It is customary to consider that values of refractive- index are for the D lines of sodium unless otherwise stated. As an example of the variation of refractive-in- dex with wave-length, an ordinary silicate crown glass had a refractive-index of 1.504 for 770m[x and this steadily in- creased to 1.56 for 276mpi. 82 ULTRAVIOLET RADIATION Clear colorless glasses will transmit usually approxi- mately 90 per cent of the perpendicularly incident visible radiation but this depends upon the refractive-index. However, it is not uncommon for glasses to show a gradual decrease in transparency from one end of the spectrum to the other. This slight selective absorption may usually be predicted by observing the color of a considerable thickness of the glass. It is not uncommon to note a decrease in transparency toward the ultraviolet. Inasmuch as these characteristics are so dependent upon the constituents of glasses and the purity of the sand, etc., it is not profitable to give much data pertaining to specific glasses. Hovestadt * has presented many details concern- ing Jena glasses and Grebe 2 has described the Jena " light- filters." In recent years certain glass manufacturers of this country have given increasing attention to " optical " glasses and special glass filters. The lead glass of which many incandescent-lamp bulbs are made is highly transparent to about 350mpi but the absorption rapidly increases from this point, becoming practically total at SOOmji. A spectrogram of the near ultraviolet after passing through the " hard " glass of which some bulbs are now being made, extends slightly further into the short-wave region than the spectrogram through lead glass. The transparency of " pyrex " ex- tends slightly further than the so-called " hard " glass, ending at a wave-length about 290mpi for a specimen a few mm. thick. Borosilicate crown glass 2 mm. thick is opaque beyond 297mpi. A specimen of this kind of glass one cm. thick transmitted 8 per cent at 309m|A. Common soda glass 2 to 3 mm. thick is practically opaque beyond 330mji but a thin layer, about 0.2 mm. thick, sometimes transmits feebly nearly to 250mji. Zschimmer 3 has published various papers pertaining to the transparency of glass for ultraviolet radiation in TRANSPARENCY OF GLASSES $3 relation to chemical composition. According to him, a specimen of " uviol " glass developed by Schott (Jena) transmits 50 per cent of radiation of wave-length 305m^ when the thickness of. the glass in 1 cm. For a thickness of 1 mm. 50 per cent of radiation of wave-length 280mji is transmitted. A specimen of " uviol " crown 2 mm. thick transmitted as far into the ultraviolet as 280mji and the same thickness of " uviol " flint was transparent as far as 285mpi. A sheet of " uviol " glass as thin as a microscope cover-glass was transparent as far as 248m(i whereas an ordinary cover glass is opaque beyond 280m^. This glass is extremely useful and has been used considerably for cells and other parts of optical systems. It is said that photographs of stars through this glass reveal many more than through ordinary glass. Zschimmer 4 has also studied the influence of chemical constitution upon the transparency of glass to ultraviolet radiation. He states that boric oxide and silica when quite pure are transparent to radiation even shorter than 200mpi in wave-length. The addition of metallic oxides increases the absorption and decreases the spectral range of trans- parency. He says that sodium oxide decreases the trans- parency more than potassium oxide and lead oxide even more so. Lyman 5 has shown that boric oxide is not as transparent as quartz in the ultraviolet for equal thick- nesses. Fritsch G has published a recipe for a durable glass which he claims is transparent to ultraviolet radiation as far as 185mji. It specifies 6 grams of commercial calcium fluoride mixed with 14 grams of boric oxide, both in powdered form. This is melted in a platinum crucible and the liquid is poured out on an unheated sheet of platinum, taking care to avoid too rapid cooling. Pfliiger 7 measured by means of a thermocouple the transparency of various glasses as far as 357mji. Kruss 8 determined the spectral transmission-factors of glasses to 84 ULTRAVIOLET RADIATION about SOOmp, by means of a fluorescent-screen photometer. Pfliiger determined the absorption-factors of various clear glasses, 1 cm. in thickness for the near ultraviolet. His results, presented in Table XVII, show the spectral transmission characteristics of various kinds of clear glasses. He used a thermopile in making his determina- tions and his original article may be of interest to some in this connection. It is apparent from these data that the heavier glasses are less transparent in the near ultra- violet region than the lighter crown glasses. A similar result is indicated in Table XVI. TABLE XVII Absorption-factors (in per cent) of Various Clear Glasses of 1 cm. Thickness Wave-length, mju 357 388 415 442 500 640 Borosilicate crown 4.7 2.5 1.2 0.7 0.5 Calcium-silicate crown Heaviest baryta crown. . . . Telescope flint 3.4 35.0 49.0 2.5 9.8 30.0 1.8 5.2 12.0 1.4 3.4 3.6 0.5 2.5 0.7 0.3 1.6 0.7 Baryta light flint.. 9.0 6.0 2.7 1.6 Baryta light flint Silicate flint 18.0 28.0 8.6 9.6 2.5 4.1 2.1 0.9 0.6 Heavy silicate flint 41.0 28.0 6.9 0.9 0.5 The spectral transmission characteristics, the transmis- sion factors and the colors of glasses are affected by tem- perature. A number of colored glasses have been studied by the author 10 and the transmission-factors determined for the heterogeneous light from a tungsten lamp. The temperature range was C to 350 C. In general, the transmission-factors decreased with increase in tempera- ture, the decrease being as great in one case (a copper red glass) as 58 per cent for the whole range in tempera- ture. In one case the transparency increased with the TRANSPARENCY OF GLASSES 85 temperature. The coloring ingredients in the glasses were copper, gold, manganese, cobalt, chromium, and others. In some cases the color-change with increasing temperature indicated a decrease in transparency to the radiation of shorter wave-lengths. In other cases the reverse was indicated. Apparently there is a shift in the absorption-band accompanying a change in temperature. Cobalt blue glasses showed little change in transparency. There is an excellent field of research open in this con- nection. What the influence is upon the transparency to ultraviolet radiation is unknown. Gibson X1 studied the effect of temperature upon the coefficient of absorption of certain glasses of known com- position. The coloring elements in the glasses were cadmium, selenium, and uranium and the temperature range was from -180 to 430 C. They found an enor- mous increase in the absorption-factor in certain parts of the spectrum as the temperature increased. In the case of selenium red glass, for example, the absorption-factor for certain wave-lengths was about 50 times greater at 430 C than at -180 C. The absorption-bands shifted toward longer wave-lengths with increasing temperature but inasmuch as the colors of their glasses ranged only from canary to red, no general conclusion can be drawn for glasses of all colors. A systematic investigation of greater extent should yield much of interest. Apparently no studies have been made upon the effect of temperature upon the spectral transmission in the ultraviolet region. In some cases the results for the visible region encourage one to conjecture concerning the ultraviolet region but this is seldom a safe procedure, especially without a broad acquaintance with spectral characteristics. It has been observed for years that arc-lamp globes gradually acquire a purplish tinge. This has been studied spectrally and there is no doubt that the purplish tint is 86 ULTRAVIOLET RADIATION due to manganese. 12 The color is readily driven out by the application of moderate heat and inasmuch as this temperature is below that of the softening point of glass, the author has suggested that glassware which has as- sumed this purplish tint can be cleared up by heating in a suitable oven. The practice of introducing manganese into clear glass which is to be used outdoors merely to counteract the blue-green tint common to iron impurity in the silica is open to criticism where a slight blue-green tinge is not objectionable. The manganese increases the absorption of the glass and by counteracting the tinge due to iron, really makes a light shade of " smoke " glass. For lighting glassware outdoors and for many sky- lights the almost unavoidable slight tinge of blue-green is not objectionable especially when the transparency is not only initially reduced by manganese but is decreased more and more on exposure to radiation. Some arc-lamp globes which were examined showed a reduction in trans- parency to 55 per cent of their initial value. 12 There is strong evidence that the " bringing out " of the purplish manganese color is due to the short-wave radiation near the spectral limit of transmission of glass. For example, glass globes in which tungsten lamps have burned for years show no purplish tint but the same globes exposed to sunlight will develop the color. Ultraviolet radiation, cathode rays, and X-rays bring out various colors in glasses and various crystals. F. Giesel 13 used a quartz mercury arc on various glasses. Out of eight glasses, four showed a change in color within 15 minutes which developed to deep violet in 12 hours. Giesel states that the coloring is due to manganous sili- cate. Gortner 14 studied the action of solar radiation on several pieces of glass and found that glass containing as much as 0.2 per cent of manganese assumed the purplish tinge of manganese in a few weeks. The depth of color- TRANSPARENCY OF GLASSES 87 ing increased with the time of exposure but some pieces containing manganese did not discolor. It appears that the development of the purplish tinge is not solely a matter of the presence of manganese but it is likely to be dependent more or less on chemical state, conditions of manufacture, or the ingredients associated in the glass. Rontgen radiation is said to cause a re- ducing action in some cases and ultraviolet produces its effect by oxidation. Oxides of iron, manganese, and chromium are the most common coloring elements of minerals. Blue uviol (Jena) glass 2 mm. thick transmits radiation as far into the ultraviolet as 285mji. The author 15 noted the interesting fact that, by the incorporation of a slight amount of cobalt into a certain glass (the constitution otherwise remaining the same), the limit of transparency was extended slightly further toward the shorter wave-lengths. Perhaps this partially accounts for the transparency of blue uviol beyond SOOmji. Didymium imparts a very complex spectral transmission curve to glass. When dense enough a didymium glass will absorb the mercury yellow lines (578m|i) without seriously diminishing the intensity of the green line (546mpi). Its spectral transmission curve shows many maxima and minima but the chief one is at about 580mpi and extends very markedly between 570 and 590mji. Neodymium colors glass a purple and praseodymium imparts a greenish yellow color. Filters for isolating the other mercury lines and for other purposes are discussed elsewhere. 16 Most of the elements which impart a yellow, orange, or red color to glass effectively reduce the transparency for the near ultraviolet. An exception is gold red which in its lighter densities is a rose-purple, fairly transparent to the near ultraviolet. Cerium in glass is of value in absorbing the ultraviolet. It imparts only a slight color. 88 ULTRAVIOLET RADIATION Chromium in glass strongly absorbs the ultraviolet. Quantities of less than 1 per cent are opaque to ultra- violet when the glass is 1 or 2 mm. thick. Cobalt (blue) and nickel (blue) and manganese (purple) have little or no effect in the near ultraviolet region. Copper (blue-green) and iron (blue-green) have only a slight effect in this region. Lead reduces the extent of the transparency of glass in the ultraviolet. Uranium in glass in amounts of 1 to 4 per cent effec- tively absorbs ultraviolet radiation. Crookes 17 made and examined spectroscopically over 300 tinted glasses of known composition with the aim of producing a glass which effectively absorbed the invisible radiation at both ends of the visible spectrum. By super- posing on the radiation from a Nernst lamp, the radiation from a high-tension electric discharge between poles of pure metallic uranium he claimed to have a practically continuous spectrum from 200mji to SOOmpi. His speci- mens of glass were of the form of flat plates 2 mm. thick. For measuring the amount of infra-red obstructed by the glass specimen he used a thermopile and a plate of biotite (black mica) which transmits nearly all the infra-red while absorbing 80 to 90 per cent of the light. A plate of dark smoky quartz accomplishes the same result as biotite but Crookes found it less easy to work with than the latter. He also employed a radiometer balance. A standard flux was used in all his mixtures. He obtained glasses whose short-wave limit of trans- parency varied from the ordinary limit for glass to the short-wave region of the visible spectrum. Several of them absorbed nearly all the infra-red radiation from his source. They varied in transparency to light over wide limits. The commercial Crookes glass is one of the best specimens. TRANSPARENCY OF GLASSES 89 The author 16 has published the spectral transmission curves of many coloring media in glass. These refer only to the visible spectrum but they are often of value in considering the ultraviolet region. There are so many commercial glasses of indefinite composition that it appears futile to present spectral data. Some data of this character is to be found in the refer- ences already given. Recently Coblentz and Emerson 18 have presented spectral transmission curves for many commercial glasses, particularly for the visible and infra- red regions. Gibson and McNicholas 19 and also with Tyndall 20 have also presented spectral data pertaining to commercial glasses. The compositions of the glasses are not given but the form of the spectral transmission curve in the case of colored glasses containing one chief coloring metal, is sufficient for one familiar with the relation of color to composition of glasses. By comparing the curves given in the papers mentioned above with the spectral curves presented elsewhere by the author, 16 the principal coloring ingredients may be determined in many cases. The spectral transmission curves of cobalt blue glasses have a sharp maximum in the violet and also a sharp rise in the deep red. Those for moderate and light den- sities also exhibit a characteristic secondary maximum at 560m^. Smoke glasses contain a variety of coloring ingredients, the aim being that the combined effect is one of complete neutralization of color. This is seldom achieved because " smoke " glasses are usually bluish, reddish, or purplish. The presence of cobalt in these glasses is often distinguish- able by the presence of the secondary maximum at 560mpi in the spectral transmission curve. Uranium (fluorescent) glass often exhibits an irregular spectral transmission curve, the location and character of the maxima and minima indicating the presence of uranium in glasses of unknown composition. It com- 90 ULTRAVIOLET RADIATION monly shows an absorption-band at about 410m| J i and a transmission-band at about 370m^. The spectral transmission curves of amber glasses differ considerably in form depending upon the ingre- dients. The presence of a sharp absorption-band at SSOmpi and the shape of this band in the spectral transmission curves of Crookes glass indicates the presence of didymium or a near relative. The approximate short-wave limits of transmission for a number of commercial glasses which are sufficiently standardized or described to be of value, are given in Table XVIII. The data have been obtained from various references already given. TABLE xvm Approximate Short-wave Limit of Transparency of Various Glasses Trade Name Thickness Limit Ultra mm. 0.39 m/z 280 Pyrex 0.77 300 Nultra 4.84 360 Noviol A 2.00 410 Noviol B 3.22 440 Noviol C . 4.23 460 Crookes A 1.79 345 Crookes B 1.93 350 Luxf el 2.00 340 Rifleite 3.14 475 Euphos 1.95 400 Fieuzal 2.13 350 Akopos 2.17 410 Hallauer .... 1.90 410 Chlorophile 1.98 350 Saniweld (light) 1.82 500 TRANSPARENCY OF GLASSES 91 " Ultra " (Corning) glass is not appreciably colored and is fairly transparent to SOOmjj,. A thin specimen transmits to about 260mji but one 0.4 mm. thick becomes opaque at 280mpi. This glass transmits further into the ultra- violet than most glasses. " Nultra " (Corning) glass is yellowish and absorbs to some extent between 500 and 400m( J i but a specimen a few mm. thick is practically opaque at the short-wave limit of the visible region; that is, at about 390 or 400mji. " Noviol " (Corning) glass is a yellowish glass which has a high transparency for the visible spectrum but absorbs the ultraviolet completely when 2 mm. thick. It selectively absorbs blue and violet light. Euphos glass 21 quite effectively absorbs the ultra- violet radiation although in thin layers it still transmits slightly throughout the near ultraviolet. Coblentz and Emerson 18 have presented the trans- mission-factors of a large number of commercial glasses for the total radiation from four sources, namely, a gas- filled tungsten lamp, a quartz mercury arc, a magnetite arc, and the sun. (See Chapter VIII). References 1. Jena glass, 1902. 2. Zeit. f. Inst., 21, 1901, 101. 3. Zeit. f. Inst. 23, 1903, 360. 4. Zeit. f. Electrochem. n, 1905, 629. 5. Astrophys. Jour. 28, 1908. 6. Phys. Zeit. 8, 1907, 518. 7. Phys. Zeit. 4, 1903, 429. 8. Zeit. f. Inst. 23, 1903, 197 and 229. 9. Ann. d. Phys. u, 1903, 561. 10. J. Frank. Inst. 187, 1919, 225; J. Amer. Cer. Soc. 2, 1919, 743; Color and Its Applications, 1921, 396. n. Phys. Rev. 7. 1916, 194. 12. Gen. Elec. Rev. 20, 1917, 671. 92 ULTRAVIOLET RADIATION 13. Elec. 1905, 1053. 14. Amer. Chem. J. 39, 1908, 157. 15. J. Frank Inst. 186, 1918, in; Color and Its Appli- cations, 1921, 399. 16. Color and Its Applications, 1921. 17. Phil. Trans. Roy. Soc. London, 214, 1914, i. 18. Bur. Stds. Tech. Pap. No. 93. 19. Bur. Stds. Tech. Pap. No. 119. 20. Bur. Stds. Tech. Pap. No. 148. 21. Trans. I. E. S. 9, 1914, 472. mju 'gs 435.8 || "3 03 g 407.8 f| 404.7 |" O o> ! > ^ 366.3 ^E 365.5 1 365.0 as | 334.2 g llJ 3/3.2 a 3/2.6 I| 302.4 2! 2 296.7 5|| 289.4 ! | 280.4 3 I' CHAPTER VII REFLECTION OF ULTRAVIOLET RADIATION In general, it may be said that ultraviolet radiation undergoes considerable absorption by reflection. Of course, this is to some extent true of any radiation, but substances which possess reflection-factors of high values for ultraviolet radiation do not appear to be as common as the so-called " white " substances from the viewpoint of visible radiation. Many substances reflect the near ultraviolet fairly well but the middle and extreme ultra- violet are generally absorbed to a considerable degree. Much of the ultraviolet energy is absorbed by some sub- stances in the excitation of fluorescence. Snow is an excellent reflector of the ultraviolet energy in solar radiation. Photographs of the spectrum of solar radiation reflected from snow show that the spectrum extends practically undiminished to 295mpi which is ap- proximately the limit of the solar spectrum. These spectrograms indicated that the ultraviolet radiation is almost totally reflected as is the case for visible radia- tion. The reflection-factor of snow for visible radiation is commonly between 80 and 90 per cent. The high reflection-factor of snow for ultraviolet radiation is one factor contributing toward the development of snow- blindness. The tremendous quantity of energy reflected into the eyes from an unusual angle and the relatively greater transparency of the lower eyelids are very im- portant factors. There are many occasions where the reflection-factors of substances should be investigated. In photography and in other photo-chemical activities at least the near 93 94 ULTRAVIOLET RADIATION ultraviolet must be reflected. It is common to use merely a so-called " white " pigment or other substance, however, there is a great difference in the reflection- factors of " white " substances for ultraviolet radiation. For example, zinc white, sometimes called Chinese white, will photograph darker than some equally white sub- stances upon which some of it has been placed. This indicates a lower reflection-factor for the radiation af- fecting ordinary photographic emulsions than the other white surface. This is only one example of many, but there are so many kinds of surfaces, substances, and combinations, that it does not appear worth while to attempt to present data. It is a simple matter to determine qualitatively the spectral reflection characteristic of any surface by means of a quartz spectrograph or even two quartz lenses and a quartz prism. If the ordinary spectrograph is used it will be found advantageous to cut a hole in the side of the instrument near the point at which it makes an acute angle with the plate-holder. The surface to be investigated should be placed in the position usually occupied by the photographic plate or the fluorescent focussing screen. On looking through the hole which has been provided one sees the visible portion of the spectrum of the source and will often see the ultraviolet spectrum faintly fluorescent because most substances fluoresce at least slightly. Now if the eye is replaced by a camera at the hole, the spectrum of the source, reflected by the substance under investigation, can be photo- graphed. By such a procedure it is almost as easy to determine qualitatively spectral reflection characteristics as it is to ascertain spectral transmission. Hagen and Rubens 1 studied the absorption of ultra- violet, visible and infra-red radiation by thin layers of metals. They 2 also investigated the reflection-factors of metals. They employed a direct photometric method ULTRAVIOLET RADIATION REFLECTION 95 using an arc as a source and worked as far into the ultra- violet as 25 Om^. Some of their results are condensed in the paragraphs which follow. Gold possesses high reflection-factors for the long-wave visible radiation but there is a minimum in the near ultra- violet between 350 and 400mji. It reflects 39 per cent of radiation of wave-length 25 Omjj,. The reflection characteristic of copper is similar to that of gold. Its reflection-factor is about 26 per cent in the region of 250m[i. The reflection-factor of platinum is fairly high in the visible region, sloping downward toward the ultraviolet and is about 34 per cent at 250m[i. Iron and nickel appear to resemble platinum in reflec- tion characteristics. Iron reflects 33 per cent and nickel 38 per cent of radiation of wave-length 25Um[i. A new silver mirror, of course, reflects very well in the visible. Its reflection-factor was found to drop to 4 per cent at 316m^ and then to rise to about 34 per cent at Mach's magnalium consisting of 69 parts of aluminum and 31 parts of magnesium reflected more than 80 per cent throughout the visible and dropped only to 67 per cent at 250mji,. This is an unusually high reflection-factor for the middle ultraviolet. The permanence of its surface is said to be excellent. Various alloys exhibit high reflection-factors in the visible region which generally decrease with decrease in wave-length in the ultraviolet. Ross's alloy, consisting of 68.2 parts copper and 31.8 parts tin, reflects about 56 per cent at 400mpi and about 30 per cent at 25 Om^. Brandes-Schiinemann's alloy, consisting of 41 parts copper, 26 nickel, 25 tin, 8 iron, 1 antimony, reflects about 50 per cent at 400m^, and about 36 per cent at 250m|i. Schroder's alloy, 66 copper, 22 tin, and 12 zinc, reflects 96 ULTRAVIOLET RADIATION about 60 per cent at 400mpi and 40 per cent at 250mji. Another Schroder alloy, 60 copper, 30 tin, and 10 silver, has reflecting characteristics similar to the preceding one. According to Lyman 3 the Brashear alloy of which his grating was constituted and with which he did such excel- lent work in the extreme ultraviolet, resembles the Ross alloy in spectral reflection characteristic. He states that it appears likely that the reflection curve for Ross alloy suffers a minimum in the middle ultraviolet and then rises again in the region of extremely short wave-lengths. He sounds a caution to the effect that it is not safe to predict the behavior of metals in the Schumann region from data obtained on the less refrangible side of wave- length 250m|i < . Voigt 4 measured the relative phase retardation and the ratio of amplitudes of the two components vibrating at right angles to each other, involved in the polarization of radiation reflected from metallic surfaces. From these data the refractive-index, the absorption-coefficient, and the reflection-factor for normal incidence may be com- puted. Minor 5 used such a method in studying the char- acteristics of metals in respect to ultraviolet radiation and obtained values of reflection-factors for radiation of various wave-lengths. For details, the reader should refer to the original paper. His computed values of re- flection-factors (in per cent) for perpendicular incidence are given in Table XIX for various polished metals. Minor's results for steel are in fair agreement with those of Hagen and Rubens. 6 In fact, his values for the various metals agree as well as might be expected with the other data available. In general, the differences are greatest for radiation of the shorter wave-lengths. Silver shows a sharp band between SOOmij, and 330mpi, the minimum being at about 315m\i. Minor presents in his original paper more detailed results for silver than is ULTRAVIOLET RADIATION REFLECTION 97 TABLE XIX Reflection-factors of Metals (in per cent) m/j Steel Cobalt Copper Silver 226.5 34.8 18.4 231.3 35.7 31.8 29.0 19.9 267.3 39.6 39.7 27.9 24.1 298.1 42.6 45.7 26.4 15.4 316.0 .... .... .... 4.2 325.5 44.8 .... 8.5 346.7 51.1 31.5 68.0 361.1 51.2 .... .... 77.4 395.0 53.5 57.7 40.1 87.1 450.0 55.4 63.3 60.5 91.7 500.0 56.9 65.6 55.5 93.2 550.0 57.7 66.6 58.4 94.2 589.3 58.4 67.5 74.1 95.0 630.0 59.0 80.5 given in Table XIX. He also measured the optical con- stants of steel, cobalt, copper, and silver. Steel and co- balt exhibit anomalous dispersion, their dispersion curves being very similar to each other, but the refrac- tive-index curve of steel has a weak minimum at 326mjj,. Copper exhibits normal dispersion but the refractive- index changes more or less abruptly at SOOmpi and at 550mjx. For shorter wave-lengths than 250m^ the dis- persion of copper again becomes anomalous. The dis- persion of silver is anomalous between 225mn and 280mjx, normal between 280mpi and SQSmji, and in the visible spectrum it becomes anomalous. According to Minor the maximum refractive-index of silver is 1.57 (at 295mjx) and the minimum is 0.155 (at SQSmji). The refractive- index of steel steadily increased from 1.3 (a,t 226mpi) to 2.65 (at 630m|i). That of cobalt steadily increased from 1.1 (at 231mp,) to 2.15 (at 590mjj,) and that of copper decreased from about 1.4 (at 235mpi) to 0.562 (at 630mpi). 98 ULTRAVIOLET RADIATION Meier, 7 employing the same method as Voigt and Minor, investigated the optical constants of bismuth, gold, iodine, iron, mercury, nickel, platinum, selenium, zinc, and two alloys. His results compare favorably with the re- sults of others but close agreement cannot be expected because of the very probable differences in the surfaces as to polish, etc. A summary of Meier's computations of the spectral reflection-factors for various metals is pre- sented in Table XX. In his paper Meier compares his results with those of Rubens and Hagen, Pfliiger, Shea, Kundt, and others. His values for gold are in general somewhat greater than those of Rubens and Hagen but they compare favorably in the case of nickel. The copper- silver alloy consisted of equal parts of the two constitu- ents. The composition of Wood's alloy is not given but it will be noted that the reflection-factors for the ultra- violet are quite great. Wood's alloy is easily made of bismuth, lead, tin, and cadmium. An alloy of these metals consisting respectively of 15, 8, 4 and 3 parts, melts in the neighborhood of 75 C. Possessing as it does high reflec- tion-factors for ultraviolet radiation, this alloy or modi- fications of it may be useful. The mercury as used by Meier was contained in a glass cell but corrections were applied so that the results presented are for a mercury- air surface. Frehafer 8 investigated the reflection and transmission of ultraviolet radiation by sodium and potassium for a spectral range from 250mpi to 550m^. She used the quartz mercury arc as a source and a sodium photo-electric cell as the recording apparatus. For almost normal in- cidence (10 deg.) the reflection-factor of sodium was high through the range studied. It was about 80 per cent from 250mjA to 400m^, then it rose to 88 per cent at 450mpi, and it remained at about 92 per cent from this point to 550mjx. On the other hand, the reflection-factor of potas- sium was about 96 per cent at 550m|i and rapidly decreased ULTRAVIOLET RADIATION REFLECTION 99 TABLE XX Reflection-factors (in per cent) of Various Metals mfji Gold Nickel Plati- num Bis- muth Zinc Iodine Alloy cu+ag Wood's alloy Mer- cury Sele- nium 257.3 27.6 30.7 37.1 20.1 20.6 16.2 52.7 23.3 274.9 27.5 37.6 43.1 24.8 47.6 16.0 66.6 25.3 298.1 30.4 39.4 47.6 31.2 60.2 .... 19.3 61.1 31.8 325.5 36.1 40.4 48.9 36.0 68.2 15.0 27.6 64.9 65.7 32.5 361.1 37.7 41.2 52.4 42.5 70.5 26.0 36.6 65.2 70.6 30.3 398.2 39.4 50.6 57.5 46.7 71.6 30.0 44.0 68.8 73.1 30.5 441.3 42.3 56.1 58.4 48.9 73.2 33.3 62.5 67.6 74.2 29.2 467.8 43.2 59.6 58.9 50.8 74.3 34.2 68.6 66.1 74.7 28.4 5G8.0 57.4 62.1 58.9 62.2 75.1 34.0 66.6 68.6 74.6 27.2 589.3 81.5 65.5 59.0 54.3 74.5 30.3 81.3 70.1 76.3 25.1 668.0 88.3 68.3 59.4 57.2 73.1 88.1 70.6 23.4 to about 12 per cent at 25 Om^. These two metals repre- sent respectively perhaps the highest and lowest reflec- tion-factors of any metals for ultraviolet radiation. They were studied in the form of opaque mirrors in contact with fused quartz plates. The transmission-factors of thin films of these two metals decreased steadily from 250mpi to 550mjA. Frehafer also investigated the ratio of the reflection-factors (at 45 deg. incidence) for light polarized with the electric vector respectively parallel and perpen- dicular to the plane of incidence. This ratio was found to possess a maximum value near 334mjx for both sodium and potassium but was greater for the former. Nutting 9 determined the reflection-factors of steel, cyanine, selenium, and glass. He employed a photo- graphic polarization photometer and sparks of aluminum, zinc, and cadmium for the ultraviolet region. Some of his results are presented in Table XXI for radiation of various wave-lengths. 100 ULTRAVIOLET RADIATION TABLE XXI Reflection-factors (in per cent) of Various Substances m // Cfppl Col pri 11 m f^vfl tiinp Glass III jut otcci OClclllU-lIl \*t jr cl 11111 C A B C 257 33.3 7.83 274 35.3 10.0 6.8 4.67 7.48 9.55 309 43.0 .... 6.8 4.62 6.94 .... 334 43.7 12.7 6.9 4.48 6.58 9.48 361 48.8 6.6 .... 395 .... 15.1 5.7 4.31 6.86 9.4 410 52.0 430 16.8 4.0 470 54.0 16.7 2.0 4.2 6.6 .... 510 16.9 1.7 .... 9.0 534 65.4 650 18.0 9.2 9.3 589 56.9 18.4 13.3 4.16 5.37 9.39 620 14.3 13.8 The results obtained by Nutting for the steel mirror show a rapid decrease in reflection-factor for radiation of shorter wave-length than 400mjj,; that is, the reflection- factor is fairly high (above 50 per cent) throughout the visible spectrum but it decreases rapidly in the ultra- violet region. Selenium mirrors were made by fusing the element on glass and covering it with plate glass, removing the latter when cool. The data show an increase in reflection- factor from the red to the yellow-green and a gradual drop toward the ultraviolet region. Cyanine exhibits a strong narrow absorption-band in the yellow and therefore is an interesting substance for the study of anomalous dispersion of solids.. Mirrors of it were made by casting. The data presented in Table XXI are Nutting's values for a fresh surface. The reflec- ULTRAVIOLET RADIATION REFLECTION 101 tion-factor rises in the red region to a maximum in the yellow. It falls to the low value of 1.26 per cent in the green (SOOmjx) and then gradually increases to a nearly constant value in the ultraviolet region. Nutting noted that exposure to radiation produces great changes in the reflection-factor and other optical con- stants of cyanine. Among other changes, the anomalous dispersion disappeared after the surface had been exposed for some time to radiation. This exposure did not affect the reflection-factors for ultraviolet radiation. The glasses whose reflection-factors were determined by Nutting were " telescope crown " designated in Table XXI as A; baryt-flint, B; and C, a glass of high refractive- index (1.9) and low dispersion. The reflection-factors are for a single polished surface in each case. E. O. Hulburt 10 has presented data pertaining to the reflection-factors of various metals and alloys for ultra- violet radiation. He used a concave reflecting grating of speculum metal and was able to record wave-lengths from 380 to lOOmji. For a source of ultraviolet radiation he employed an end-on hydrogen tube as devised by Lyman. 3 The tube was of the internal capillary type equipped with a fluorite window and filled with hydrogen at about 1.5 mm. pressure of mercury. This was excited by a 11 00- volt transformer. For measuring the radiation he used a sodium photo-electric cell connected to an elec- trometer. Most of the data obtained by Hulburt are presented by him in the form of curves from which the values in the following descriptions were taken. He found that the reflection-factors for radiation of shorter wave- lengths than SOOmfx are rarely above 50 per cent although silicon was a noteworthy exception for it showed a re- flection-factor of 76 per cent in the region from 200 to SOOmpi. He obtained brilliant opaque films of silicon by cathodic sputtering. In most cases he polished the 102 ULTRAVIOLET RADIATION metals with rouge and chamois. Some data pertaining to the metals studied by Hulburt are as follows: Aluminum. Films were made by cathodic sputter- ing on glass from a freshly scraped aluminum cathode in mercury vapor. The reflection-factor varied from about 70 per cent at 380m^ to about 23 per cent at 180 m\jL. Antimony. A polished cathodic film showed reflec- tion-factors varying from about 32 per cent at 380mpi to about 14 per cent at ISOmjx. Bismuth. A polished cathodic film steadily decreased in reflection-factor from about 37 per cent at 380 m^ to 15 per cent at 180m[i. Cadmium. A polished cathodic film varied in reflec- tion-factor from about 60 per cent at 380mjA to about 20 per cent at ISOm^. Carbon. A cathodic film showed very low reflection- factors compared with silicon, which is its neighbor in the periodic system and which has similar electrical proper- ties. Its reflection-factor remained practically constant at about 16 per cent down to about 210mji and then grad- ually decreased to 10 per cent at ISOm^. Carborundum. A perfect surface of a crystal showed reflection-factors increasing from about 11 per cent at 3SQm\i to about 20 per cent at 180mji. The curve is of the same character as that for any dielectric, such as quartz. The value of the reflection-factor indicates a very high value of refractive index. Chromium. Both a solid polished plate and polished cathodic films were used. The polished plate showed reflection-factors decreasing from 30 per cent at 380m|i to about 18 per cent at 320mpi then remaining at this value to about 25 Om^. A defined minimum of about 16 per cent was found at 240mpi and a maximum of about 23 per cent at 215mji. The reflection-factor then steadily decreased to 18 per cent at 180mji. ULTRAVIOLET RADIATION REFLECTION 103 Cobalt. A polished sheet of rolled metal showed a steady decrease in reflection-factor from 55 per cent at 380mpi to about 33 per cent at ISOmjj,. The values throughout the region studied were comparatively high. Copper. Both electrolytic and cathodic films were used. The former showed reflection-factors decreasing from 35 per cent at 380mp, to 25 per cent at 250mpi with a rise to 32 per cent at 210m^i and a decrease to about 20 per cent at ISOmji. The cathodic film showed gen- erally higher values throughout. The mean value was about 32 per cent from 380 to 200mji. Gold. Cathodic films showed a decided maximum of 35 per cent at about SOOmpi, the reflection-factor de- creasing to 15 per cent at ISOmpi. A polished plate gave practically the same results as old cathodic films. Be- tween 300 and 380m[x the reflection-factor was about 30 per cent. Lead. A cathodic film decreased in reflection-factor from about 42 per cent at 380m|i to about 21 per cent at ISOmji. Magnalium. A good mirror of this alloy of aluminum (69 parts) and magnesium (31 parts) was not produced by Hulburt, but such as it was it showed a decrease from about 47 per cent at 380m^ to 9 per cent at ISOmpi. Hagen and Rubens obtained values from 81 to 67 per cent between 357 and 251mjj, respectively. Magnesium. A polished mirror showed a steady de- crease from 48 to 25 per cent from 380 to 220mji respec- tively thence a more rapid decrease to 11 per cent at Molybdenum. A polished specimen showed reflection- factors of about 40 per cent from 380 to 300mpi; thence a rapid decrease to a marked minimum of 24 per cent at 250mji. The value was about 30 per cent between 230 and 200mj J i after which it again decreased. Nickel. A polished electrolytic film plated upon a 104 ULTRAVIOLET RADIATION cathodic film showed a decrease from 50 per cent at 380m|A to 37 per cent at 250mji, thence to a defined maxi- mum of 46 per cent at 211m^. The reflection-factor then decreased rapidly to 30 per cent at ISOmpi. Palladium. A cathodic film showed a very gradual decrease from 30 per cent at 380mji to about 15 per cent at 180m|A. Selenium. A mirror was made by pouring melted selenium upon glass and removing it after it became cold. The reflection-factor decreased from 30 per cent at 380mpi to 10 per cent at ISOmji. Silicon. A piece of silicon was ground with emery and then polished with rouge. It and other specimens showed very high values of reflection-factors, namely 60 per cent or greater throughout the whole range from 380 to ISOmjx. Silver. Two specimens showed decided minimums at about 310mpu The reflection-factor decreased rapidly from about 70 per cent at 380mpi to about 5 per cent at 310mpi. The values then increased to about 33 per cent at 230mjx and then decreased to about 16 per cent at Speculum. This alloy of copper (68 parts) and tin (32 parts) is used for making reflection gratings. All of Hulburt's curves showed a gradual decrease in reflec- tion-factor with decreasing wave-length. The mirror giving the highest values varied in reflection-factor from more than 60 per cent at 380m^ to 21 per cent at 180mjx. This was a freshly polished surface. Exposure to air gradually reduced the reflection-factor for all wave- lengths. Steel. A polished piece of hardened steel showed a decrease in reflection-factor from 55 per cent at SSOmjx to 17 per cent at 180m^. Stellite. A polished specimen of this alloy of chro- mium and cobalt exhibited a maximum (42 per cent) at ULTRAVIOLET RADIATION REFLECTION 105 215mpi and a minimum (36 per cent) at 240mj4,. Chro- mium exhibits a similar maximum and minimum. The reflection-factors for the stellite mirror decreased from about 60 per cent at SSOmji to 30 per cent at Tantalum. A polished specimen showed a gradual decrease in reflection-factor from 28 per cent at 380mp, to 12 per cent at ISOm^,. Tellurium. A polished cathodic film decreased in reflection-factor from 40 per cent at 380m[i to 23 per cent at 250mji thence the decrease was very gradual to 20 per cent at ISOmpi. Tin. A cathodic film three days after polishing showed reflection-factors decreasing from 31 per cent at 380m^i to 7 per cent at ISOm^. Tungsten. A polished specimen showed a reflection- factor of about 30 per cent from 380 to 310mpi. It then gradually decreased to a value of about 15 per cent at 260m^. The value then remained practically constant to 180m|A. Zinc. A polished cathodic film showed a reflection- factor of about 51 per cent from 380 to 300m[x. It then gradually decreased to 15 per cent at ISOmji. The data obtained by Hulburt shows that throughout the region from 380 to ISOmpi the reflection-factor de- creased in general with decreasing wave-length and was never zero. The high reflection-factors of silicon indi- cate that this metal would have advantages for mirrors and gratings in ultraviolet investigations and should have other practical applications. Platinum and nickel also appear to be especially serviceable in the ultraviolet. 106 ULTRAVIOLET RADIATION References 1. Ann. d. Phys. 8, 1902, i. 2. Ann. d. Phys. n, 1903, 873. 3. Astrophys. Jour. 23, 1906, 181. 4. Phys. Zeit. 2, 1901, 303. 5. Ann. d. Phys. 10, 1903, 581. 6. Zeit. f. Inst. 19, 1899, 293; 22, 1902, 52. 7. Ann. d. Phys. 31, 1910, 1017. 8. Phys. Rev. 15, 1920, no. 9. Phys. Rev. 16, 1903, 129. 10. Astrophys. Jour. 42, 1915, 205. Relative Quartz Mercury Arc Exposure Carbon Arc 32 n n 16 u ii 8 u n 4 n II 2 n II / Iron Arc J2 " n 16 // n 6 n H 4 n n 2 Quartz Mercury Arc 16 II n 8 II II // 4 II it " 2 n H " / Plate IV. Ultraviolet spectra of the ordinary carbon arc, the iron arc, and the quartz mercury arc as obtained by various photographic exposures. The exposures for one radiant are not directly comparable with those of the other radiants. CHAPTER VIII ULTRAVIOLET RADIATION IN COMMON ILLUMINANTS There are many sources of ultraviolet radiation among which to choose in meeting any particular conditions; however, there are few which are powerful enough to be widely applicable to industrial arts or even to replace solar radiation in photo-chemistry. The ideal source for many purposes, especially in scientific investigation, is one emitting a continuous spectrum of high and uniform intensity but this does not exist for any large range in the ultraviolet regions. It is a common practice to use a spark which emits many spectral lines in combination with a continuous-spectrum source when absorption spectra are to be determined. The principal sources available are discharge tubes, sparks, arcs, incandescent solids, and combinations of these. If the source must be surrounded by a container, as in the case of the discharge tube and the mercury vapor lamp, the spectral character of the radiation is im- mediately limited by the spectral transmission character- istic of the container. Thus glass limits the spectral range of radiation from the common mercury (glass- tube) lamp and quartz does in the case of the quartz mercury lamp. The transmission- and reflection-factors of materials used in connection with sources of ultraviolet is of primary importance. Discharge tubes equipped with fluorite windows of the best quality have enabled investigators to explore the ultraviolet spectrum beyond the transparency limits of quartz and other media but fluorite becomes opaque at about 120mpi. At this point it must be abandoned and the source must be placed in 107 108 ULTRAVIOLET RADIATION the same vessel as the photographic plate or other re- cording device. Even the vessel must be evacuated owing to the absorption by air. The sun is still a much-used source of radiation when only the near ultraviolet is required. Notwithstanding its unreliability, it is depended upon for various purposes owing to the high intensity and to the relatively low cost of its radiation compared with artificial radiation for photo-chemical activities. However its supremacy in photo-chemical arts is being seriously menaced in various quarters. In Tables XXIV and XXV it is seen that solar radiation is not extremely rich in ultraviolet energy as compared with artificial sources. It is its overwhelming intensity which makes sunlight relatively so effective. Incidentally solar radiation ranks higher than any other illuminant in luminous efficiency, that is, in lumens per watt. From radiation and illumination measure- ments luminous efficiency may be computed. Ives, 1 using data obtained by Kimball, 2 computed the luminous efficiency of solar radiation at the earth's surface. His results are given in Table XXII. TABLE XXII Luminous Efficiency of Solar Radiation Lumens per Direct sunlight watt Zenith distance 48.3 " " 66.5. " " 73.5. Blue-sky radiation Overcast-sky radiation. 86.5 78.9 71.1 137.3 96.5 On comparing the values in Table XXII with the luminous efficiencies of artificial illuminants in Table IN COMMON ILLUMINANTS 109 XXIV, it is seen that the latter are much smaller. The excessively high luminous efficiency of solar radiation is due primarily to the high temperature of the sun but also to the absorption of infra-red radiation by the atmos- phere. This means that a great portion of solar radiation reaching the earth's surface is in the visible region. The limelight which was produced by Drummond about a century ago was one of the earliest artificial sources of radiation which emitted an appreciable quantity of near ultraviolet energy. This consists essentially of a piece of rare-earth oxide (zirconia, etc.) in a hot flame such as oxyhydrogen. This has the advantage of burning in the air without a container and therefore has some uses in experimental work. However the Nernst glower is more convenient and usually more satisfactory. The Nernst glower, now practically obsolete as a com- mercial illuminant, owes its efficiency to rare-earths and high melting-point. It has the advantage of operating in air without an enclosure. Allen 3 states that its spectrum extends to 210mji but is very faint between that wave- length and 25 Om^. Its greatest photographic effect (ordinary emulsions) is in the violet and blue regions. Many of the early experiments on photo-chemistry were done with burning magnesium usually in the form of ribbon. This source finds uses even at the present time despite its inconvenience and unsteadiness. It has the advantage of requiring no equipment except a match to ignite it. The gas-mantle, which was a great improvement in the production of artificial light, emits little ultraviolet radia- tion except that close to the visible spectrum. It owes its efficiency to rare-earths. Eder 4 investigated the photo-chemical intensity of some of the earlier sources. Of course, the term " chemical intensity " is very indefinite and the results depend upon the particular reaction used. In this case the action on 110 ULTRAVIOLET RADIATION silver bromide was employed as a mode of comparing the various sources. His results are presented in Table XXIII. The results with the hefner lamp, burning amyl acetate, are taken as a standard and the " relative acti- nism" is the ratio of the chemical effect upon silver bromide of the radiation from one source upon silver bromide to that of the hefner lamp. These data serve only to show very approximately the relative amounts of violet and ultraviolet radiation emitted in each case. In each case the source of radiation was at a distance of one meter from the silver bromide. TABLE xxm Relative Effects Upon Silver Bromide Relative Intensity Relative Actinism Visual Chemical Hefner .... 1 70 16 60 400 135 1 260 28 160 4000 435 270-400 769 1 3.7 1.75 2.6 10.0 23.8 Limelight. . Argand Welsbach.. Arc Magnesium Magnesium Magnesium Magnesium per 1 mg. in air through clear glass in oxvcen Jones, Hodgson, and Huse 5 made an extensive study of the photographic efficiencies of various common illumi- nants. They employed the three classes of photographic emulsions, namely, ordinary, orthochromatic, and pan- chromatic. Some of their data are presented in Table XXIV. In the second column are the values of luminous efficiency which aids in specifying the conditions of opera- tion of the sources. The values in the last three columns pertain to the relative photographic action per watt, the sun's action on the three emulsions being taken as 100 in IN COMMON ILLUMINANTS 111 each case. These values are based upon energy-con- sumption. There are many cases where it is of interest to have values upon a visual basis ; that is, to know the photo- graphic (or other action) per unit to brightness or per foot-candle. A table of these values is found in the original work but they may be obtained in any case from Table XXIV by dividing the photographic efficiency by the lumens per watt and multiplying the quotient by 150 in order to obtain values in terms of those for the sun. This procedure gives what may be termed relative actinic value per lumen. TABLE XXIV Relative Photographic Action Per Watt for Three Classes of Emulsions Source Lumens per watt Relative Photographic Efficiency Ordinary Orthochromatic Panchromatic 1 Sun 150 100.00 100.00 100.00 2 Acetylene 0.7 0.14 0.21 0.24 3. Acetylene screened 0.07 0.04 0.04 0.04 4. Pentane 0.45 0.05 0.9 0.13 Mercury arc 6. Fused quartz 40 158.0 130.0 99.0 6. "Nultra" glass... 35 50.0 47.0 39.0 7. Crown glass 37 79.0 68.0 62.0 Carbon arc 8. Ordinary glass 12 10.0 9.0 8.5 9. White flame 29 52.0 45.0 42.0 10. Enclosed 9 11.0 11.0 10.0 11. "Aristo" 12 62.0 86.0 60.0 12. Magnetite arc 18 12.0 14.0 10.0 Incandescent filament 13. Carbon 2.44 0.37 0.52 0.68 14. Carbon 3.16 0.51 0.74 0.95 15. Tungsten, vacuum 8.0 1.7 2.2 2.7 16. Tungsten, vacuum 9.9 2.4 3.0 3.5 17. Tungsten, gas-filled 16.6 6.1 6.8 7.7 18. Tungsten, gas-filled 21.6 8.9 9.8 11.0 19. Mercury vapor tube 23.0 47.0 54.0 42.0 112 ULTRAVIOLET RADIATION The plates used in obtaining the data in Table XXIV were as follows: Ordinary. Seed 23, chiefly sensitive from about 360mjo, to about SOOmji. Orthochromatic. A special experimental plate, chiefly sensitive from about 400mji to about 610m[i. Panchromatic. Wratten, chiefly from about 400m[A to about 720mjA. The data thus gives an idea of the relative quantities of radiation of these spectral ranges emitted by these various sources. The data in Table XXIV is very useful in judging com- mon illuminants as to " actinic " value but the descriptive notes which follow must be considered in connection with them. In some cases it will be noted that a glass lens was interposed between the source and the photographic plates. A few notes pertaining to the sources included in Table XXIV are as follows: 1. The sunlight exposures were made on a clear day between 1 :30 and 2 :30 p. m. 2. The acetylene burner was standard, the flame be- ing cylindrical. 3. The acetylene source was also screened with a blue filter that transmitted light closely approximating average daylight. 4. The pentane source was a standard Harcourt lamp adjusted according to standard specifications. 5. The quartz mercury arc operated at 220 volts and 3.4 amperes. A reflector consisting of a polished plate of black glass 2 cm. thick was employed, it being assumed that the reflection from the surface was non-selective. The intensity was reduced by a pair of quartz lenses. 6. This source was also screened with a piece of heavy lead glass 4 mm. thick, known as " Nultra " ; IN COMMON ILLUMINANTS 113 It was quite colorless and had a transmission- factor of about 90 per cent for the visible rays. 7. Conditions were the same as in 6 except that one of the quartz lenses was replaced by one of clear crown glass. 8. The carbon arc was an automatic-feed type with carbons at right-angles. It operated at 110 volts d. c. and 6 amperes with a 60-volt drop across the arc. The positive carbon was 6 mm. in diameter and cored. The intensity was reduced by glass lenses. 9. The white flame arc operated at 115 volts d. c. and 24 to 26 amperes with an 85-volt drop across the arc. The lower white flame carbon was 10 mm. and positive; the upper carbon was 13 mm. and cored. The flame was 2.5 to 3 cm. long. The intensity was reduced by means of one quartz and one crown glass lens. 10. The enclosed arc was enclosed by a glass cylinder with close-fitting metal ends. The carbons were of ordinary cored type, at right angles, and the positive crater was fully exposed to the photo- graphic plate. It operated at 110 volts d. c. and 8 amperes with a 65 -volt drop across the arc. 11. The "Aristo" was an enclosed arc with vertical carbons the positive being above. It operated at 220 volts and 16 amperes. The length of the arc was about 2.5 cm. 12. The magnetite arc was of the commercial type. It operated at 110 volts d. c. and 4 amperes. Its intensity was reduced by glass lenses. 19. The mercury vapor arc was in a glass tube 45 cm. long and 2.8 cm. in diameter. It operated at 115 volts d. c. and 3.5 amperes, the actual drop across the tube being 3.5 amperes. Radiation from a section of the tube 2 cm. long in the middle of the length was tested. 114 ULTRAVIOLET RADIATION The values in the table for tungsten lamps are appli- cable only to those particular lamps. The operating efficiencies of incandescent lamps are gradually increasing and they are determined by lighting considerations. The economic factors of photo-chemical arts are quite different. For example, in ordinary photography the exposure is usually short and the operating temperature of the fila- ment of a tungsten lamp used for this purpose can be con- siderably above the normal for ordinary lighting without introducing a prohibitive expense due to the short life of the lamps. A blue-glass bulb was developed by the author several years ago in order to have a photographic incandescent lamp of high actinic value per lumen for such fields as portraiture. For the same filament-tempera- ture its actinic value per watt is about the same as a clear- glass lamp of the same size and its actinic value per lumen is about three times as great. These values are based upon the ordinary photographic plate. However, this tungsten photographic lamp is designed to operate at a higher filament-temperature than a clear-glass lamp of the same size and therefore is more powerful photographically. The values for the blue-bulb photographic tungsten lamp presented by Jones, Hodgson, and Huse are very obviously not for the standardized blue-bulb photographic lamp and not for the proper operating conditions. This lamp is useful in many cases where only a moderate amount of near ultraviolet radiation is required. The radiation of short wave-lengths from a tungsten filament increases in quantity much more rapidly with increase in filament- temperature than the radiation of long wave-lengths. For this reason it is sometimes well to operate the tungsten lamp at voltages considerably above normal. It has been found 6 that, for an ordinary photographic plate (Seed 23) and a 1000-watt gas-filled tungsten lamp operating at 18 lumens per watt, an increase of 17 per cent in voltage above normal doubles the photo- IN COMMON ILLUMINANTS 115 graphic action. In other words, for the ordinary plate an increase in voltage from 115 to 135 volts reduced the watt- age to one-half in order to obtain the same photographic action. For an increase of 17 per cent in voltage the photographic action on an orthochromatic plate increased 67 per cent. The extent of the spectra of some common illuminants and a brief discussion of them has been presented else- where by the author. 7 Spectral transmission-curves are also given for lead and for " Euphos " glasses. The spec- tra presented are for a mercury arc (glass tube) skylight, a tungsten (vacuum) lamp, a quartz mercury arc, a yellow flame arc (opal glass), a carbon arc (bare) and a carbon arc (clear glass). The magnetite arc exhibits a spectrum of titanium and iron which is extremely rich in lines throughout the near ultraviolet region and down to 230mji. It is especially rich between 330 and 290mpi, a region of particular interest owing to the fact that this is the region where the solar spectrum ends. Bell 8 has presented data pertaining to the ultraviolet energy in artificial light-sources. He used a Rubens thermopile to receive the radiation from the source. This was connected with a sensitive galvanometer and the radiation from a standard carbon incandescent lamp was used as a standard. It was found that 1mm. on the gal- vanometer-scale equalled 35.3 ergs per second per square centimeter at the thermopile. He used a quartz cell con- taining a thickness of 1cm. of distilled water. The quartz and distilled water are transparent to ultraviolet, visible, and infra-red radiation from about 120 to 1300mp,. The ultraviolet radiation was separated from the remainder by using a piece of " Euphos " glass of proper thickness and quality. The procedure was to measure the quan- tities of radiation reaching the thermopile through the cell, with and without the Euphos in the path. Bell's 116 ULTRAVIOLET RADIATION TABLE XXV Ultraviolet Radiation from Various Sources Deflections due to ultra- violet (cm.) Total ultra- violet ergs per sec. per sq. cm. Ultraviolet ergs per sec. per sq. cm. per ft-cand. Quartz Hg. arc, opal glass Graetzin gas lamp 3.70 0.92 1305 4.3 11.7 Carbon (gem) filament, 100-watt Cooper Hewitt Hg. tube, glass. . . Sunlight direct 0.61 1.64 2.28 215 577 14.8 15.6 16.1 Acetylene flame. . . . 0.52 18.4 Tungsten, vacuum, 100-watt Nernst lamp, glass globe Magnetite arc, glass 1.90 1.81 22.40 670 640 7900 22.7 25.5 30.3 Magnetite arc, quartz Quartz Hg. Arc, bare, old Quartz Hg. arc, bare, new Carbon arc, quartz 29.00 16.77 32.10 74.00 10240 5920 11350 26200 36.3 38.3 87.6 91.0 results are presented in Table XXV. In the second column are given the galvanometer deflections due to the ultraviolet radiations. These are given in centimeters so that these values must be multiplied by 353 to give the actual ergs per second per square centimeter due to ultra- violet radiation from the various sources. The energy values are found in the third column and afford a direct comparison of the relative amounts of ultraviolet energy supplied by the various sources. They represent the absolute amounts of energy received per second per square centimeter at a distance of 50 cm. from the sources. In the last column are the amounts of ultraviolet energy ac- companying the energy necessary to be appraised visually as one foot-candle. It is interesting to note the decrease in output of ultraviolet radiation as the age of a quartz mercury arc increases. This perhaps has been noted by those who Plate V. The white flame arc a powerful source of *' near radiation approximating solar radiation in many of its effects. ultraviolet IN COMMON ILLUMINANTS 117 have employed the quartz arc for its ultraviolet energy. Bell's data indicate a very great difference between the old and the new quartz mercury arcs which he examined. The new one emitted more than twice the amount of ultraviolet emitted by the old one notwithstanding they were rated equally in watts. The magnetite arc was of commercial type and it emitted as much ultraviolet radiation as the new quartz mercury arc. The electrical input was perhaps about the same in the two cases. The carbon arc was of the enclosed type operating at 6.6 amperes d. c. A quartz window was placed in the globe and it was found that 30 per cent of the energy emitted through the quartz window was cut off by the Euphos glass. There are powerful bands in its spectrum between 350 and 360m^ and between 380 and 390mpi. A compara- tively small change in the output is effected by interposing clear glass. Verhoeff and Bell 9 have made an extensive investi- gation of the pathological effects of radiant energy and the eye. This will be discussed in a later chapter. The data in Table XXVI is taken from this work and it may help to convey an idea of the absorption of various media. The transmission-factors of the various media are for radiation from a commercial magnetite arc. They found that this arc operating at 9 amperes and 750 watts pro- vided at a distance of 50 cm. an intensity of ultraviolet energy shorter than 390mj,i in wave-length, equal to 15000 ergs per second per sq. cm. of which about 3500 ergs was of shorter wave-length than 300m jj, as against 57.00 ergs per second per sq. cm. for the quartz mercury arc at the same distance in the same spectral region. The transmission-factors for the total radiation from four common light-sources have been determined by Coblentz and Emerson 10 for a large number of commercial glasses. A few of these are presented in Table XXVII. Presumably the three artificial light-sources were operated 118 ULTRAVIOLET RADIATION TABLE XXVI Transmission-factors for Radiation from a Magnetite Arc Filter Transmission-factor Two quartz plates each 3 mm. thick 53 per cent Same plus 5 cm. distilled water (water-cell) 33 Water-cell and dense flint glass (limit 335mju) 26 Water-cell and medium flint glass (limit 315mju) 28 Water-cell and light flint glass (limit 305mju) 27 Water-cell and crown glass (limit 295mju) 28 Dense flint (n = 1.69) alone 40 Medium flint (n - 1.63) alone 45 Light flint (n - 1.57) alone 40 Crown (n = 1.51) alone 43 at their commercial ratings. The quartz mercury arcs were new. The tungsten lamp was a 500-watt gas-filled stereopticon lamp. There are various types of mercury arcs available but they may be divided into two chief classes those with glass-tubes and those with quartz-tubes. The glass-tube mercury arc or a quartz mercury arc equipped with a glass globe, emits ultraviolet radiation in the near region, that is, of wave-lengths longer than about SOOmjj,. The quartz arcs emit in general the near and middle ultraviolet. They are made in large sizes and also in various " concen- trated " forms but their radiation does not differ materi- ally. The spectrum of the quartz mercury arc has many powerful lines but also a number of gaps. Between the powerful group of lines at 365m^ and the double line at 313mji there is only one strong group and that is at 334mpi. A gap exists between 313m^ and a group at 302.5mji and another between the latter and 297mji. The output of mercury arcs operating at high tem- peratures as is the case of the quartz arc diminishes con- siderably as the lamps age. It has been found that the ultraviolet spectrum of the quartz mercury arc dimin- ishes in intensity as the lamp ages, especially for radia- IN COMMON ILLUMINANTS 119 TABLE XXVII Transmission-factors of Various Media for Radiation from Four Different Sources Medium Thickness in mm. Transmission-Factor Tungsten lamp Quartz Hg. Arc Magnetite Arc Sun Freuzal B, green-yellow. . . Euphos B, green-yellow. . . Akopos, green-yellow Noviweld 6, green-yellow Saniweld, dark, amber. . . Noviol B, yellow. 2.04 3.12 1.68 2.17 1.32 2.88 1.97 2.00 71.6 78.8 84.6 0.9 78.1 74.1 85.3 75.7 2.6 67.8 26.9 24.7 29.5 0.4 10.6 32.2 46.1 32.0 7.2 7.9 59.6 64.9 35.4 43.1 56.0 83.0 46.0 63.0 59.0 0.2 43 66 64 1.2 48 63 64 74 0.9 60 76 89 69 12 48 82 92 76 Crookes A. neutral Crookes B, neutral Gold plate, film Selenium, red 2.90 1.85 1.66 1.30 0.09 10.00 10.00 10.00 Window glass, clear Crown glass Mica, brown Mica, colorless Water, clear 34.2 Clear water (glass cell) . . . Clear water (quartz cell) . . tion shorter than 254mji,. It has been suggested that this decrease may be due to the greyish deposit on the walls of the tube or that the composition of the gas undergoes change. It may be well to try cooling the quartz tube in an attempt to preserve the constancy of the output of ultraviolet radiation. This decrease in ultraviolet intensity with age has been verified by physical, chem- ical, and biological tests. It has been stated that the average life of the 220-volt quartz mercury arc is between 2500 and 3000 hours and that the tubes can be re-exhausted about three times. A great deal of work has been done in investigating 120 ULTRAVIOLET RADIATION the spectrum of mercury. Huff 11 has discussed the spectra of mercury as obtained from mercury arcs in air and in tubes and also in the spark. He found that by increasing the capacity in the secondary of a coil giving an alternating current discharge, the spectrum changed from that of the arc to one containing the characteristic lines of the spark. The introduction of self-induction in the secondary of such a coil tends to reduce the spectrum of the spark to that of the arc. The arc in this case con- sisted chiefly of a hollow carbon rod filled with mercury. He found the arc-spectrum obtained in this manner to extend into the short-wave region as far as 187mjA when photographed for one hour by means of a large grating spectro graph. Arons 12 gives a list of the more intense lines of the quartz mercury arc. There are a large number of lines emitted from the arc in a quartz tube, extending from 230m^ to 579m^. Line 254mpi is very intense and, ac- cording to Hughes, is well situated so as to be very effec- tive in producing photo-electrons from most metals. There are a number of weak lines between 185m[A and 200mji. These do not appear strong on a photographic plate but they are quite active photo-electrically. The quartz mercury arc emits lines of still shorter wave-length which are absorbed by the fused quartz tube. Hughes 13 placed a mercury arc and a metal plate in the same vacuum. By assuming Ladenburg's law the velocity of emitted photo-electrons increases with decreas- ing wave-length and is proportional to the frequency of the exciting radiation Hughes concluded that the mer- cury spectrum extends to 123mpi but the radiation is weak between 145mpi and 178mji. He showed that the mercury arc in a fused-quartz tube whose walls were 0.5 to 1 mm. thick, emitted radiation as short as 184.9m| > i in wave- length. Lyman observed line I77.5m\ji from a fused quartz arc. IN COMMON ILLUMINANTS 121 Anthracene is photo- electrically active under the radia- tion from a quartz mercury arc but it is not photo-electric to radiation greater than 220mj J i in wave-length. There- fore radiation of shorter wave-length than 220mpi is trans- mitted by the fused-quartz tube. The transparency of air must be reckoned with in this region of the spectrum. For example, the mercury line, 184.96mpi, will be quite effective when the air-path is short but will be feeble or may be completely absorbed if the air-path is long. Thus the transparency of fused quartz, and finally of air, limits the extent of the spectrum of the quartz mercury arc under ordinary conditions. Inasmuch as there is a gap in the spectrum of the mercury arc between 190m[i and a point near 185mpi, the former wave-length is likely to be the limit of the spectrum under most conditions. Considering the diminution of the ultraviolet spectrum due to aging of the quartz mercury arc it seems reasonable to set 200m[4, as the usual limit of the spectrum of the quartz mercury arc. Of course the limit of the glass-tube mercury arc is determined by the glass and, therefore, the limit of the spectrum of the glass mercury tube is near SOOni^. Hallwach 14 measured the energy of the various mercury lines by a thermo-electric method and obtained the rela- tive values in Table XXVIII. For the sake of future reference his values of relative photo-electric activity of these lines for potassium are also included. The relative values in Table XXVIII are referred to those for 436m^ in each case. Of course, the relative intensities of the various lines vary considerably depend- ing upon various factors, but those in this table give an TABLE XXVIII Wave-length, m/* 678 546 436 406 365 313 254 217 Relative energy 116 169 100 67 119 90 38 55 Photo-electric activity (potassium) Relative actual 3.2 8.3 100 79 218 301 198 390 Relative specific. ... 2,7 4.9 100 118 183 334 520 710 122 ULTRAVIOLET RADIATION idea of what may be expected. The photo-electric effects of the various lines upon potassium are referred to as " relative actual." By correcting these values to a uni- form energy (dividing values in third line by those in the second line) the relative specific photo-electric activities of radiations of various wave-lengths are obtained for potassium. Lyman 15 studied the spark-spectrum of mercury as far into the extreme ultraviolet as 126mj4,. He gives a table of about 100 lines between this wave-length and 205mji' for the spark spectrum. For the arc-spectrum the short- est wave-length in his table is about 140mji and he pre- sents only 5 lines between this and 185m^i. An indication of the emission of short-wave radiation of high intensity by the quartz mercury arc is the ozone with which the air is charged in the vicinity of one of these sources. Incidentally it is well to wear glasses when working with this lamp and various bare arcs to prevent the painful irritation in the outer eye-media which is a common result of middle ultraviolet radiation. The lines in the spectrum of the mercury arc are often quite readily isolated by means of filters so that they afford intense sources of monochromatic or homogeneous radiation. The two yellow lines 577 and 579m|i (represented ap- proximately by 578ml!) can be isolated by a yellow filter having a steep absorption-band on the short-wave side so that line 546m[x is absorbed. Chrysoidine and cosine are fairly satisfactory if the former is diluted and the cosine is added until the green line disappears. An aqueous solution of neodymium ammonium nitrate or neodymium chloride (or a neodymium or didymium glass) of sufficient density, possesses a very sharp absorp- tion-band in the vicinity 578mjj, sufficiently strong to absorb the yellow lines and to transmit the green line, 546mji, quite freely. IN COMMON ILLUMINANTS 123 Solutions of Neptune green S and chrysoidin isolate line 546mpi fairly well but not as satisfactorily as the preceding filter. Many yellow filters will eliminate the lines of shorter- wave-length and the infra-red will be almost completely absorbed by several centimeters of water. One centi- meter of water is opaque to radiation longer than HOOmji. The opacity of water for infra-red can be increased by the addition of copper chloride. Coblentz ie has discussed filters for the infra-red. A solution of potassium bi-chromate is satisfactory for eliminating the blue, violet, and ultraviolet lines. The line 436m^ can be eliminated by using a solution of potassium permanganate and one of nickel nitrate of just sufficient strength to eliminate the other lines. The two solutions should be kept separate. Another fairly satisfactory combination for isolating this line is dense cobalt (blue) glass and a solution of quinine. A solution of aesculine absorbs the ultraviolet lines quite completely and if dense enough it may be used in the foregoing to replace the quinine for eliminating line 405m^i. Line 405mji is fairly well isolated by methyl violet and quinine sulphate in separate cells. It also transmits lines 408m^ and 398mjj,, the latter rather faintly. Lines 365 and 334mji are transmitted by methyl violet and nitrosodimethylaniline. Line 334mji, may be elimi- nated by thick glass. R. W. Wood has recommended for isolating line 313mpi, a silver film deposited upon a quartz plate, but according to Hughes, if the film is thick enough to absorb the adjacent lines it very greatly diminishes the intensity of SISmji. He found a certain sheet of mica transmitted 313m^ almost entirely but still was quite opaque to 303mpi. This provides at least a sharp cut-off on the short-wave side. 124 ULTRAVIOLET RADIATION A great variety of arcs may be devised by the investi- gator but there are only a few commercial arc-lamps available. These are the common carbon arc, open or enclosed, the various flame arcs which utilize carbons impregnated with various chemical salts and the mag- netite arc consisting of a positive electrode of copper and a negative one of magnetite (iron oxide). The carbon and especially the white flame arcs are in use extensively for photo-engraving, dye-testing, etc. They can be arranged for a wide variety of wattages. The magnetite arc is in use largely for street-lighting but it can be adapted to many other purposes owing to its fairly intense ultraviolet radiation. Most commercial arcs are designed to consume less than 1000 watts but by in- creasing the size of the electrodes much more powerful sources can be devised. The ultraviolet spectrum of the ordinary carbon lamp is confined chiefly to the near ultraviolet, although there are some strong lines and many weak ones in the middle region. Relatively more ultraviolet radiation is emitted by the faint flame of the arc than by the crater from which most of the light comes. By impregnating the carbons with various elements and salts a wide variety of spectra can be obtained; for example, by using iron compounds a rich iron spectrum may be obtained ex- tending throughout the near and middle regions to 200mji. It is the iron oxide which makes the magnetite arc rich in ultraviolet radiation. Spectra of the mercury, iron and carbon arcs will be found elsewhere. 17 Lindemann 18 studied the radiation from the carbon arc by observing the photo-electric effect of a copper-oxide sensitive plate. He found the greatest effect due to the radiation from the violet tufts at the extremities of the electrodes which consist of unburnt carbon vapor. He also concluded that impregnated carbons produce smaller photo-electric effects than ordinary carbons. IN COMMON ILLUMINANTS 125 The impregnation of carbons with various elements and compounds has been a fruitful field of research. It has resulted in the development of carbons for so-called " flame arcs." In these more light, as a rule, is emitted by the arc (flame) than by the craters. The introduction of various materials into carbons results in quite a variety of colors, and Mott 19 has done a great deal of work in this field in recent years. Calcium fluoride produces yellow light which consists really of green and red lines superposed on a more or less white light. The yellow flame arc emits moderate amounts of violet and ultra- violet radiation. Strontium fluoride is chiefly respon- sible for the color of the red flame arc. Various other colors are produced by the following materials: copper, blue ; silicon or iron, red-violet ; titanium, blue ; didymium oxide, violet; thorium oxide, reddish; eerie oxide, blue; lanthanum oxide, blue. Of all the flame arcs the one in widest use at the present time, owing chiefly to its extreme richness in near ultraviolet radiation, is the white flame arc. It owes its spectral qualities largely to rare earths. Ac- cording to Mott and Bedford 20 "the radiation of the snow-white flame arc is the closest approach to sunlight plus blue sky of any known illuminant, considering either the visible or the ultraviolet spectrum. The spectrum is a mass of lines crowded close together and is confined chiefly to wave-lengths longer than SOOmpi. The ultra- violet radiation of the snow-white flame arc does not produce ozone in quantities detectable by the odor as in the case of the quartz mercury arc, indicating the absence of the short wave-lengths which produce ozone." Ac- cording to Mott there is a marked decrease in photo- graphic effect in the region of SOOni^ compared with other flame carbons or the open arc. Flame carbons have been made which emit much energy between 200 and SOOmjLi. 126 ULTRAVIOLET RADIATION Mott and Bedford 20 have presented comparative results obtained with the various flame arcs for several chemical reactions. These present better than a detailed descrip- tion some of the relative merits. In Table XXIX are their results pertaining to the effect of different sources on a 10 per cent solution (by volume) of bromine in toluol contained in glass and in quartz. The sources were at a distance of two feet and the time required for the bromine color to disappear was measured in each case. TABLE XXIX Bromination Time for Different Sources Glass Tube Quartz Flask White flame arc, 26 amp. 90 arc volts 36 36 Yellow flame arc, 25 amp. 90 arc volts. 25 24 Red flame arc, 25 amp. 90 arc volts 170 21 Blue flame arc, 25 amp. 90 arc volts 210 60 Pure carbon arc 280 58 Tungsten lamp, 760-watt, gas-filled clear bulb 274 Tungsten lamp, 1000-watt, gas-filled blue bulb 235 Mercury vapor lamp, 3.3 amp. 110-volt 610 The last three sources in Table XXIX are enclosed with glass so that the results obtained with them are more directly comparable with those obtained with the arcs when the bromine was contained in the glass test tube. The results of a test of the effect upon solio paper of the radiations from the flame arcs operated at 25 amperes and 90 arc-volts are presented in Table XXX. The recip- rocals of the average time required in each case for equal coloration is the basis of the figures. The glass used in this case was ordinary window glass 2.3 mm. thick. Paraphenylenediamine and nitric acid provide a reaction which is a satisfactory test for ultraviolet radiation, and Plate VI. The quartz mercury arc shown with the quartz arc exposed and also as used for exposing materials to its radiation. IN COMMON ILLUMINANTS 127 TABLE XXX Relative Photographic Effect on Solio Paper No Glass With Glass White flame . . . 100 80 Yellow flame 35 17 Red flame 30 22 Blue flame 60 30 Pure carbon arc 40 24 Mott and Bedford used it for comparing the various arcs. They impregnated white blotting paper with a solution, consisting of 1 gram of paraphenylenediamine, 3 cc. dis- tilled water, 2 cc. nitric acid (sp. gr. 1.21). The blotting paper was then dried in a steam oven, for if it is dried, slowly as by leaving it over night it blackens and is use- less. This is not appreciably sensitive to the radiations from the gas mantle, electric incandescent filament lamps, and the Nernst glower. On exposure to ultraviolet the impregnated paper turns green and finally to a metallic brown color. Mott and Bedford found that best test results were obtained at a distance of two feet from the 25 ampere arc on exposing in increments of one-half minute without glass. Owing to the absorption of glass for the ultraviolet the exposures when it was used were increased to five minutes. Their results are presented in Table XXXI. This gives a rough estimate of the relative amounts of ultraviolet radiation referred to the blue flame arc. Without glass the blue flame arc was the most powerful but with glass the white flame arc was quite superior. The same investigators compared the relative efficacies of the radiations from the white flame arc and the quartz mercury arc of approximately the same wattage, in pro- ducing chlorination. The radiation from the white flame 128 ULTRAVIOLET RADIATION TABLE XXXI Relative Photographic Effect on Paraphenylenediamine No Glass With Glass Ratio White flame . . . 50 12.0 4 to 1 Yellow flame 40 0.8 50 to 1 Red flame 30 0.15 200 to 1 Blue flame 100 4.0 25 to 1 Pure carbon arc 20 1.0 20 to 1 arc appeared to be superior from every viewpoint. A decided advantage possessed by the white flame arc is that it is very efficient even when glass vessels are used. Furthermore it can be made to emit enormous quantities of radiation if necessary. Mott 21 made elaborate tests of the flame arc in paint and dye-testing and concluded that at a distance of two feet it provided an intensity more intense than summer sunlight. Inasmuch as its spectrum is similar to day- light on a clear day and as it is reproducible and control- lable it should find greater usefulness. Certainly where powerful sources of near ultraviolet and other radiations which are chemically active are needed, the white flame arc is perhaps the most promising commercial source available. Of course the quartz mercury arc possesses some advantages in steadiness and operation over long periods without any attention. Various other photo-engraving arcs are available but they are usually ordinary carbon arcs of high current- density. Henri 22 concluded that the intensity of the ultraviolet radiation from a quartz mercury arc increases with in- crease in temperature of the tube. By cooling the tube with running water he found the intensity of the ultra- violet to be only one-fourteenth as powerful as when the tube was surrounded by air. IN COMMON ILLUMINANTS 129 The commercial gas-filled tungsten lamps emit a con- tinuous spectrum of appreciable intensity in the near ultra- violet. By operating them at excessive voltages the in- tensity of the ultraviolet can be greatly increased. For some work it is advantageous to cement a quartz window to such a lamp. Gehlhoff 23 has presented some data pertaining to the distribution of energy from tungsten and tantalum spirals immersed in nitrogen or argon. He equipped the lamp with a quartz window. Although the ultraviolet spec- trum is weak and is confined to the near region, this source has the advantage of being continuous. It is useful in some investigations. As has already been stated the output of ultraviolet radiation from an incandescent filament lamp increases very rapidly with increase of the temperature of the fila- ment. Furthermore the operating conditions of incan- descent filament lamps used for photochemical reactions should not be the same as those which have been stand- ardized from an economic viewpoint for such lamps when used for ordinary lighting purposes. Additional photo- graphic results as obtained by L. L. Holladay, A. H. Taylor and the author are shown in Table XXXII for four representative photographic plates. The Mazda C lamp consumed 1000 watts and operated at its normal voltage at an efficiency of 16.8 lumens per watt. The latter is a direct indication of the temperature of the fila- ment. The Mazda C photographic blue-bulb lamp which consumed 1000 watts is designed for photographic work and therefore operates at a higher filament tempera- ture. This increase in temperature over that of the normally operating filament corresponds to an increase in voltage of approximately 10 per cent. The luminous efficiency owing to the absorption of light by the blue bulb was 10.8 lumens per watt in these experi- ments. The Cooper-Hewitt (glass tube) and quartz 130 ULTRAVIOLET RADIATION TABLE XXXH Relative Photographic Power Per Lumen of Total Radiation of Various Illuminants Source Plate Relative Photographic Value Direct Once reflected Mazda C (clear) Seed 23 it it (C (C II Orthonon 14 (< II II || Panchromatic Type M { ii ii cc tt Panchromatic W and W M i. A few of moderate intensity were recorded in the visible region. Handke 18 recorded nearly 100 lines between 160 and Gold did not emit any strong lines. The line of short- est wave-length recorded was at about 190mji and quite a few moderately weak lines appeared between 190 and 230mpi. The table shows a gap between 230 and 395mpi with a few moderately intense lines in the visible. Handke found more than 100 lines between 160 and 200m(i. Tin did not yield any strong lines. Pfliiger found a group near 190m|x, a gap from that point to 210m|x, and weak lines between 210 and 313m[A. Moderately intense lines appeared throughout the near ultraviolet and visible regions. Handke recorded about 30 lines between 170 and 200mji,. EXPERIMENTAL SOURCES 143 Antimony emitted scarcely any lines excepting a few of moderate intensity in the visible. Platinum yielded lines of moderate intensity from 190mji to the end of the visible spectrum but not as many as in the case of iron or cadmium. Palladium emitted rather weak lines between 190 and 225mji, moderately intense ones up to 260m^, but only a few weak ones between this point and the long-wave end of the visible. Magnesium yielded strong lines between 275mji and the visible spectrum. Lyman and also Handke have re- corded a few lines between 170 and 200m|i. Pfliiger recorded a few weak lines for mercury between 185 and 200m^ with the strongest of the group at about 190mji. He found a gap between 200 and 220m|i; a group of moderately intense ones between 220 and 230mji; a gap between 230 and 250mji; and moderately intense ones between 260 and the red end of the visible region. The fact that Pfliiger recorded only a few mercury lines com- pared with Lyman and others who examined the spectra only qualitatively indicates that these relatively few were the strongest ones, for he used a thermopile and deter- mined relative energy. The thermopile is enormously less sensitive than the photo-electric cell or photographic plate for detecting the presence of spectral lines. How- ever, this makes Pfliiger's results especially valuable for some purposes for they represent presumably the strongest lines in the various spectra. Of course, it should be noted that the distribution of energy may vary with conditions in the electrical circuit. Pfliiger used a spark 2 mm. long and two small Leyden jars across it. Pfliiger also presented a condensed table of strongest lines and groups of lines emitted by the metals which he examined. This is reproduced in Table XXXIII, the values of galvanometer-deflections indicating the relative intensities. It should be noted in connection with the 144 ULTRAVIOLET RADIATION TABLE XXXHI Strongest Lines and Groups of Lines in the Spark Spectra of Metals rap Metal Deflection mfj. Metal Deflection 186 Al* 173 240 Co 140 Cd* 27 241 Fe 125 190 Sn* 62 249 Fe 126 193.5 Al* 58 250.2 Zn* 80 195 Hg 40 255.8 Zn* 85 199 Al* 50 257.3 Cd* 41 202.5 Zn* 225 258 Co 80 206 Zn* 280 261.4 Pb* 12 208.7 Zn* 160 274.7 Fe 125 210 Zn* 220 274.8 Cd* 49 213.8 Zn* 60 277.1 Zn* 25 214.4 Cd* 185 280 Mg* 950 219 Ni 107 285 Mg* 153 219.4 Cd* 120 293 Mg* 189 220.3 Pb* 32 309 Mg 35 221 Ni 140 326 Sn 19 Co 53 328-334 Zn 60 226.5 Cd* 170 340-346 Cd 25 231.5 Cd* 190 361 Cd 45 232 Ni 175 365 Hg 52 233 Ag 36 o An * indicates that there were lines so close together that their combined energy was measured. preceding paragraphs pertaining to Pfliiger's work on spark spectra that he only considered spark spectra for regions of longer wave-length than ISOmpu Furthermore he was unable to record weak lines by his method. This accounts for the absence of data pertaining to lines in the extreme ultraviolet region. Ross 24 has described a powerful aluminum spark which he used in photo-chemical investigations. He connected the spark-gap to the secondary of a large induction coil with a large Leyden jar in parallel. The primary circuit EXPERIMENTAL SOURCES 145 operated at 3.4 amperes and 110 volts. By placing vari- able resistors and an ammeter in the primary circuit he was able to obtain a fairly constant source of radiation. The aluminum terminals were sharpened after each observation. They were of large cross-section and ar- ranged so that they would be cooled by conduction by dishes of ice and iron plates in contact with them. He found that the rate of decomposition of the iodides was at least twice as great in the case of the aluminum spark as in the case of any other common metals which he used. The arc is another useful source of ultraviolet radia- tion. It is usually easier to devise than the spark or vacuum tube but on the other hand it is generally less satisfactory owing to its unsteadiness. The commercial arc lamps have been discussed in Chapter VIII but for many experimental purposes an arc can be constructed very simply which is more satisfactory. The ordinary carbons can be bored and filled with metals and com- pounds, or material may be placed on the tips of the elec- trodes or in a cup drilled in the end of one of them. Arc spectra differ in general from line spectra but those ele- ments which produce spark-spectra rich in ultraviolet radiation, usually emit powerful ultraviolet radiation in the arc. Electric arc carbons impregnated with iron salts emit powerful ultraviolet radiation. In fact iron is one of the best materials for electrodes. It is a simple matter to construct an arc which will emit ultraviolet energy, provided hand-control is satisfactory. An iron rod and a carbon rod may be employed successfully for the two electrodes, however, two iron rods serve the purpose very well. These terminals can be kept cool effectively by means of heavy brass or copper sleeves which may be moved along the iron rods as the latter are consumed. A particularly successful arc of this type can be made 146 ULTRAVIOLET RADIATION in an hour or so. 25 The upper pole, which is negative, may be an iron rod about J inch in diameter. This is surrounded by a movable but well-fitted solid sleeve of copper or brass about one inch in diameter. The sleeve should be turned down to a conical shape at the arc end in order not to obstruct unduly the radiation from the arc. It is set so that about J inch of the iron electrode protrudes from the conical end. The lower electrode, which is positive, may be an iron rod about % inch in di- ameter with the end of the form of a shallow cup. One electrode should be adjustable vertically. In preparing the arc a bead of molten metal is developed in the dished end of the lower electrode by striking the arc repeatedly. This bead and the dished end of the lower electrode be- come oxidized, which apparently diminishes deterioration as the arc plays between the bead and the upper electrode. The latter is well cooled by the massive copper sleeve and the arc is steadily maintained between it and the molten bead. Such an arc will operate at a fairly high current density for thirty minutes without adjustment. Arcs between copper electrodes have been successfully used and an arc between silicon terminals emits intense ultraviolet radiation accompanied by little light. Zinc, cadmium, and aluminum have been used success- fully in the crater of the positive electrode of a carbon arc. The inner vapors of a carbon arc are carbon, cyanogen, and carbon monoxide. These alone radiate ultraviolet energy but it is neither intense nor rich in spectral lines or bands. Metals can be easily introduced into the arc in the form of wire, thin rod, or powder embedded in the carbon. Mott 56 has published interesting data pertaining to the appearance of the arc when various chemicals are melted in it. The arc can be operated in an atmosphere of hydrogen, nitrogen, etc., in order to prevent oxidation. In the study EXPERIMENTAL SOURCES 147 of the arc spectra of metals this is sometimes very satis- factory. Deposit is diminished which is essential to the success of some work. There is the danger of explosion in the use of hydrogen which must be guarded against. This danger is not present when nitrogen is used and it appears more satisfactory than hydrogen in some cases in other respects. Sometimes it has proved undesirable to operate sparks in air and in these cases other gases have been employed. Varly 2r has recommended an arc between iron termi- nals in an atmosphere of hydrogen as a constant source of ultraviolet radiation. He connected the electrodes to the secondary of an induction coil with three large Leyden jars in parallel. An alternating current of 4 amperes passed through the primary coil. He claimed that the ultraviolet radiation remained practically constant in in- tensity for days if employed for ten-second periods, allow- ing rest-periods of a few minutes between each short run. The enormous beam intensities obtained by some of the modern search-lights are due to small electrodes, the hot ends of which are cooled in a blast of air or alcohol-vapor. The electrodes are rotated in some cases. Rotation has often been applied in experimental work in effort to ob- tain uniformity. It has been found that alloys of silver and cadmium work well in the arc. Such an alloy consisting of 60 per cent cadmium melts at 700 C. The arc is steady and has been kept in one position by rotating the electrodes in opposite directions. Cadmium in a quartz tube has been utilized as a source of ultraviolet radiation. An arc between tungsten electrodes and operated in an atmosphere of hydrogen or nitrogen is a source of in- tense ultraviolet radiation. Its spectrum consists of several hundred measurable lines. The tungsten arc is best operated in a sealed bulb containing argon or a mix- 148 ULTRAVIOLET RADIATION ture of argon and nitrogen at atmospheric pressure. A heating coil placed near the arc can be used for starting the arc and the coil may serve as one electrode. How- ever, the arc is perhaps more satisfactory if two small buttons of tungsten serve as electrodes with the heating coil as an auxiliary. A series of spectra of this tungsten arc for various currents has been published by the author. 28 They were obtained for the radiation emitted through a crystalline quartz window cemented to the bulb by means of sodium silicate, the seal being further pro- tected on the outside by means of Khotinsky cement. The window was cemented upon the open end of the glass tube about one inch in diameter which protruded suffi- ciently from the bulb to remain cool. The spectrum con- sists of many lines between 200 and 400m|i. Morphy and Mullard 29 have described a tungsten arc similar to the foregoing. The lamp contains a tungsten filament which also forms one electrode. The hot fila- ment ionizes the gas so that an arc is formed between it and another tungsten electrode. The bulb is made of quartz and the output of ultraviolet radiation therefore is large. Tungsten arcs such as these in glass bulbs can be pur- chased at the present time. Several attempts have been made to employ combi- nations of tungsten and mercury. The radiation from these tungsten-mercury arcs is that due to incandescent tungsten and the mercury vapor. The total radiation is quite actinic but is limited in usefulness for work requir- ing a continuous spectrum in the ultraviolet or one packed with spectral lines by the relative scarcity of mercury lines compared with iron lines, for example, obtained from the iron arc. The mercury arcs of commercial type have been dis- cussed in Chapter VIII. Small ones can be made of glass without much difficulty but there are few persons skilled EXPERIMENTAL SOURCES 149 in working quartz. Quartz lamps of various shapes for special purposes can be supplied by manufacturers. Double-walled quartz arcs have been made so that the arc can be cooled by circulating water over the inner tube. This maintains a low density of the mercury vapor which favors the production of ultraviolet radiation. The spec- trum of mercury has been discussed in other chapters. Ellis and Wells 30 have described various special mercury arcs used by investigators or supplied by manufac- turers. By way of improving the vacuum in quartz mercury arcs, von Recklinghausen 31 has suggested the use of metals which absorb nitrogen at high temperatures. Magne- sium, boron, and titanium are proposed for this purpose. He has also studied the proper distribution of the mercury and has employed a series of electrodes. 32 For sealing molybdenum and its alloys into quartz a flux has been recommended consisting of ten parts silica, one part alumina and one part boric acid. The content of silica in the flux should be increased as the quartz is approached. Knipp 33 has described in detail the construction of a quartz mercury lamp for which he claims several advan- tages such as ease of starting and control of the mercury. It is portable and it can be taken apart for cleaning or for introducing various materials into the arc. Helbronner and von Recklinghausen 34 have, devised a powerful source of ultraviolet radiation which consists of a quartz U-tube the legs of which nearly touch each other, the electrodes of mercury being therefore side by side. The tube is 14 mm. in diameter and the legs are 160 mm. long. The lamp operates on 500 volts and takes 3 am- peres with 375 volts actually across the electrodes. The candle-power perpendicular to the axis is said to be 8000. Bovie 35 has described simple quartz mercury arcs for 150 ULTRAVIOLET RADIATION photochemical investigations. He also gives detailed in- structions for making such a lamp in a variety of forms. Anyone interested in constructing a mercury arc of special form will find it helpful to consult Bovie's paper. Henning 36 determined the mean value of the expansion coefficient of quartz to be 0.00000054 per degree per unit of length for temperatures up to 1000 C. Sand 37 has described a cadmium vapor lamp com- parable to the mercury arc. The cadmium is placed in a quartz envelope and is melted by external heating before starting. The cadmium is prevented from adhering to the sides of the quartz container by the presence of powdered zirconia. Cooper Hewitt 36 has patented the use of metals such as thallium and caesium in the construction of quartz mercury arcs in order to increase the output of ultraviolet radiation. Mercury is used for the anode and the other metal serves as the cathode. A manufacturer of mercury arcs has devised a type known as the " hot-cathode " lamp. In place of the mer- cury at the anode a spiral of tungsten wire is used. A reservoir of mercury forms the cathode. The tungsten wire is sealed into the quartz tube by a graduated mixture of glass and fused quartz so that at the wire the seal is of glass. Kowalski 39 found the oscillating spark to be more efficient as a source of ultraviolet radiation than the quartz mercury arc, particularly for sterilizing water. According to his experiments only 45 to 90 watt-hours of electrical energy were necessary for sterilizing 1000 gallons of water. Of course, the material of the electrodes, the frequency of the circuit, and other conditions affect the character of the ultraviolet radiation emitted by the spark. To increase the amplitude, the induction was reduced as much as possible and capacity was introduced. He found the relative intensities from a 22 mm. gap between invar EXPERIMENTAL SOURCES 151 terminals were 1, 1.3, 1.6, and 2 respectively for frequen- cies of 50, 40, 30, and 20 sparks per second. The per- centages of radiation determined horizontally in the di- rection of the axis of a 110-volt Heraeus quartz mercury arc operating at 3.1 amperes and 90 arc- volts were: "heat" 24 per cent; visible radiation 41 per cent; ultra- violet radiation 35 per cent. From a 22 mm. spark gap with invar electrodes at 30 sparks per second and 50.5 amperes in the oscillating circuit, the percentages were: "heat" 22 per cent; visible radiation 18.6 per cent; and ultraviolet radiation 59.4 per cent. Verhoeff and Bell 40 determined the amounts of energy radiated in various spectral regions by a 220-volt 3.5-am- pere quartz mercury tube with a voltage drop of 90 volts across the tube. A water-cell was placed before the lamp so that the radiation passing through was practically all visible and ultraviolet. Under these conditions they found that 35 per cent was visible and 65 per cent was ultraviolet radiation between 200 and 400m|i. This ultra- violet radiation was equally divided between the near and middle regions, one-half being between 200 and 300mjA and the other half between 300 and 400m^. Under these conditions the energy intensity at 50 cm. from the tube was about 11000 ergs per second per sq. cm. of radiation shorter than 400ml!, in wave-length and about 5500 ergs per sec. per sq. cm. of radiation shorter than 300mpi in wave-length. According to Tian 41 he noted in experiments on the effect of ultraviolet radiation on water that the endother- mic combinations produced by radiations of 190mji in wave-length are often destroyed by radiations of other wave-lengths. For example, in making ozone it is desir- able to avoid those rays which restrict the reaction. Quartz mercury lamps must be operated at a low voltage as the total radiation increases greatly with voltage while the radiation of 190m^ in wave-length increases much more slowly. 152 ULTRAVIOLET RADIATION He devised a lamp having a quartz tube down the center of which an insulated iron wire is passed. This makes a contact with a mercury cathode at the bottom. The anode is a cylinder of iron. The advantage claimed for this lamp is that it operates on low voltage and also it can be immersed conveniently in liquids. If alternating current is to be applied to it, the anode is made double as usual, the two being separated by means of mica. Allamand 42 determined the relative amounts of energy in the principal lines of the spectrum of a mercury arc in a uviol-glass (Jena) tube. These were determined by means of a thermopile and are referred to the blue line as 100 units. The results are as follows: Wave-length, mju 578 546 436 405 362 313 Relative energy 27 73 100 56 42 17 Lehmann 43 has described a filter for ultraviolet rays consisting of blue uviol (Jena) glass filled with a solution of copper sulphate and coated outside with gelatine con- taining nitrosodimethylaniline. A quartz mercury arc or iron arc is used as the source of ultraviolet radiation and lenses of quartz concentrate the rays. The spectrum of oxidizing phosphorus has been photo- graphed in the ultraviolet region by Centnerszwer and Petrikaln. 44 They used a solution of phosphorus in paraf- fin and passed a strong current of air over it. After an exposure of 95 hours they found sharply defined lines and a band near 325mpi. The author attempted to photograph the ultraviolet spectrum of oxidizing phosphorus but was unable to obtain any photographic action in the ultra- violet with exposures of fast plates in a quartz spectro- graph as long as 75 hours. Among the most recent work in the extreme ultraviolet region is that of Millikan 45 who extended the spectrum to about 20mu, He used high potential sparks in a vacuum and succeeded in extending the known spectrum of vari- EXPERIMENTAL SOURCES 153 ous elements to the following limits: carbon, 36.05mpi; zinc, 31.73mji; iron, 27.16mjj,; silver, 26mpi; nickel, 20.2m|A. Evidence is presented in this work which indi- cates that the characteristic L series of X-rays of carbon have now actually been obtained by ordinary mechanical gratings and that the gap between X-rays and ordinary radiation appears to have been closed. This actual linking of the extreme ultraviolet spectrum with the X-ray spectrum was accomplished by intermit- tent sparking between electrodes from 0.1 mm. to 2 mm. apart with a battery of Leyden jars charged to potentials of several hundred thousand volts by a powerful induction coil. A mercury diffusion-pump was attached to the vacuum spectrometer in order to reduce the pressure of the gases evolved by the sparking below 10 ~ 4 mm. The production of this kind of spark was found impossible if the pressure of the evolved gases exceeded the foregoing value. Specially ruled gratings were used for producing the spectra. These gratings were such as to throw as much radiation of short wave-length as possible into the first-order spectrum. Millikan, Bowen, and Sawyer 46 succeeded in producing specially ruled gratings which met the requirements of work in the extreme ultraviolet. They measured about 75 spectral lines of carbon between 36myi and 193mpi; about 200 lines due to iron between 27mj,i and 215m^i; and about 75 lines due to nickel be- tween 73mpi and 186mfi. They have presented these in their paper in the form of tables. According to Millikan 45 the substances of greatest in- terest for studies in this extreme region are those of small atomic number, for no X-ray spectra of the L series have been recorded with crystal gratings in the case of ele- ments of atomic number less than 30. Furthermore substances of atomic number much lower than this are beyond the range of the methods of X-ray spectrometry, because of the fact that the wave-lengths of the L rays 154 ULTRAVIOLET RADIATION from such substances become so large in comparison with the grating space of crystal gratings that sharp images can not be formed. The lowest limit thus far reached by Millikan was ob- tained with nickel and has a value of 20.2m[x which is between one and two octaves farther down than the lowest values previously obtained. The lowest limit found for carbon was 36.05mjA which was not at the limit of the grating. The evidence is quite convincing that Millikan's plates, obtained with the carbon spark, exhibited the whole spectrum which the carbon atom is able to emit up to and including its X-radiations of the so-called L type. Fur- thermore by examining through a quartz window the radi- ation of these high-potential carbon sparks with a fluoro- scope strong X-rays were found to be emitted. Tables have been presented by Gramont 47 which contain the ultimate lines in the dissociation spectra of 83 ele- ments. One column contains lines determined visually; another shows those obtained by means of a " crown uviol" spectrograph between 317 and 480m^i; and a third column records those below 317 m^ photographed by means of a quartz spectrograph. In another paper 48 he discusses arc-spectra of metals with low melting-point. Pierucci 49 has confirmed the conclusions of others that the spectral lines of highest excitation are confined to the center of the arc crater. He drilled an axial hole into the positive carbon and placed this carbon below and in line with the negative one. The crater was photographed through the hole by means of a reflecting prism thus eliminating the light of the outer regions of the arc. The successive elimination of low-excitation lines as the tem- perature rises was well shown by spectrograms of calcium and sodium. McLennan 50 has described a vacuum grating spectro- graph which he employed for studying the arc spectra of several elements. He provided for removing exuded EXPERIMENTAL SOURCES 155 gases, for using gratings of various sizes, and for arrang- ing its carrier and controls so that adjustments may be easily made. He made a study of a tungsten arc in helium at a pressure of 30 to 40 cm. of mercury. A small heating coil was used to start the arc and the latter could be es- tablished and maintained constant for hours with distances of 5 to 6 mm. between the electrodes. The extreme ultraviolet arc-spectra of certain metals have been studied by McLennan, Ainslie, and Fuller. 51 They employed a vacuum spectro graph with a fluorite optical system and a vacuum arc-lamp. They studied Cd, Cu, Zn, Al, C, Fe, Sn, Pb, Tl, Ni, and Co between 140 and 240mji. McLennan and colleagues 52 employed a fluorite spectrograph in the study of short-wave arc spectra in vacuo and the spark-spectra in helium of vari- ous elements. They describe the details of their ap- paratus and present tables of the vacuum arc spectra of antimony, bismuth, calcium, magnesium, selenium, silver, and copper and of the spark-spectra in helium of anti- mony, bismuth, aluminum, cadmium, lead, magnesium, thallium, and tin. The spectral region was below ISSmji. Their results with the vacuum grating spectrograph ex- tend the known vacuum-arc spectrum of copper to 122mji. They also investigated the spark-spectra of silicon, tellurium, molybdenum, and zirconium and have presented 53 a table of wave-lengths of the lines observed between 163 and 185mpi. In previous work McLennan and Lang 54 studied the spectrum of mercury down to 143.5mji, of iron to 142.7mjj,, and of carbon down to According to de la Roche 55 the spark-spectra of vari- ous elements in reducing gases may be very different in air, oxygen, carbon dioxide, and sulphur dioxide. The spark-spectra in the latter gases are very similar. The change due to reducing gases was exhibited by spectra of electrodes of Te, Mo, Ni, W, Sb, Sn, less in those of 156 ULTRAVIOLET RADIATION Ca, Ag, Au, and not at all with Zn and Cl. Self induction appeared to weaken these spectra in reducing gases. Carter and King 56 have studied the production of spectra of metals in high vacua. They vaporized Mn, Ti, Mg, and Cd by heating by means of a stream of cathode rays and excited the vapor by the bombardment of the cathode particles. In the ultraviolet region there is a relatively high intensity of lines as compared with the arc and furnace spectra. L. and E. Bloch 57 presented tables of spark-spectra containing 36 new mercury lines of rather low intensity in the region between 140 and 164mj J i. They employed a prism spectrograph and amalgams of cadmium and sodium as electrodes. They also present 18 new copper lines between 154 and 166mpi; 12 new zinc lines and 13 new thallium lines between 144 and 184mji; various lines of antimony, arsenic, bismuth and tin between 140 and L. and E. Bloch 58 have presented tables containing 67 zinc lines, 99 cadmium lines, 10 lead lines, 115 iron lines, and 143 cobalt lines in the region between HOm^ and 185mpi. Most of the lines were found to be rather weak, but two lead lines, 182.17m[x and 179.63m|i, were quite intense. Dhein 59 has published the results of measurements on the arc spectrum of cobalt. His tables contain several hundred lines between 259 and 742m| > i which are com- pared with the results of Krebs and Stiiting. He used a concave grating and specially sensitized his plates when working in the region of greater wave-length than Schumacher 60 has presented tables of wave-lengths of about 360 lines in the spectrum of the iron are in the region of 210 to 237 m^. Hicks 61 in the course of his systematic investigations of spectra has discussed the copper spectrum as well as EXPERIMENTAL SOURCES 157 that of silver, of gold, and of other elements. The arc- spectrum of copper is very rich in lines. Eder 62 has conducted a systematic investigation of the rare-earths. In this paper he presents a table of wave-lengths of about 4400 lines of the arc-spectrum of dysprosium between 228 and 700mpi. In previous papers he presents data pertaining to other rare-earths. He studied 63 a chloride or oxide of gandolium prepared from gandolium obtained by fractionation of samarin and europium. The fractions of the latter indicate spectral lines of an unknown element. Ludwig 64 has published a series of papers on spectral determinations. In this paper he discusses the arc- spectrum of vanadium between 220 and 465m[i,. He ob- tained very steady arcs operating on 0.5 amperes and 220 volts by using copper electrodes containing vanadic acid. He also used carbon electrodes impregnated with divanadyl tetrachloride. In all of his extensive work on arc-spectra the electrodes are of the respective metal or of either copper or carbon. A hole is made in the lower (positive) electrode to take the substance to be studied or the carbon is impregnated. He compares his obser- vations with those of other investigators. Belke 65 has published data pertaining to the arc-spec- trum of tungsten from 225 to 698mpi and has included comparison data. The arc-spectrum of tantalum and of molybdenum are reported on in the same volume by Josewski and by Puhl- mann respectively. The arc-spectrum of scandium has been investigated by Crookes. 66 The material was prepared from wilkite and then mixed in a powdered state with finely divided silver. This mixture was compressed into small rods which were used as electrodes. This spectrum was photographed along with that of pure silver and that of iron. A table is presented consisting of wave-lengths of 101 scandium lines from 242 to 630mji. 158 ULTRAVIOLET RADIATION The arc-spectrum of cerium nitrate was investigated by Klein 67 using a concave grating. The salt was placed in a hole in the lower (positive) carbon and an iron salt was added to obtain the iron spectrum for purposes of comparison. Tables of wave-lengths are presented for the region between 251 and 455mjx and the measurements are compared with the previous work of others. Vahle 68 using the same apparatus investigated the arc-spectrum of zirconium nitrate between 228 and Recently Hagenbach and Schumacher 69 have pre- sented tables of lines in the spectra of cadmium and of zinc observed in the electrodeless ring-discharge. The results indicate that this ring-discharge spectrum is more like the spark than the arc but that it contains more lines than the arc and spark together. The intensities in the ring-discharge spectrum are in some cases quite different from those in the arc and the spark spectra. Series in the spectrum of Argon have been discussed by Nissen 70 and numerical data are included in the paper. Burns, Meggers, and Merrill 71 have measured 55 lines in the spectrum of neon, between 336.9 and 849.5m|i, by means of the interferometer. Paschen 72 has recently presented tables containing about 850 lines of the neon spectrum from 255 to 984mpi. The infra-red lines are taken chiefly from Meissner's 7S work. The tables include intensities and frequencies and and also show to which series the various lines belong. These series are further discussed in another paper. 74 Grotrian 75 also discusses these series. The intensity relations in the spectrum of helium have been discussed at length by Merton and Nicholson. 78 They consider three factors which affect the distribution of in- tensity among the lines in the spectrum: (1) The electrical conditions of excitation; (2) the presence of impurities; (3) the pressure of gas in the discharge tube. They EXPERIMENTAL SOURCES 159 studied the effect of cathode distance, the regions of maximum emission, and the various series of lines. Lyman 77 has recently discussed the helium series in the extreme ultraviolet and Hicks 78 has added some com- ments. Compton and Lilly 79 have studied the excitation of the spectrum of helium by bombarding pure helium with electrons from a hot-filament cathode at various pressures up to 24 mm. The fact that after striking the arc it could be maintained by a potential difference as small as eight volts by using large currents indicates that in an intense discharge a large proportion of the atoms are in an abnormal state and therefore require less energy for excitation. As the voltage was increased the sharp subordinate series became relatively weaker and as the pressure increased the band spectrum became stronger and the enhanced line weaker. The band spectrum was stronger near the cathode while the enhanced line was stronger near the anode. The spectra of compound gases in vacuum tubes have been studied recently by Bair. 80 The gases were am- monia, nitrous oxide, nitrogen peroxide, carbon dioxide, hydrogen sulphide, and sulphur dioxide. The ammonia band in the visible spectrum has two heads each degraded on both sides. The band at 337.1m^ is probably not due to ammonia for it was observed in tubes long after the characteristic color of ammonia disappeared. The dis- charge-tubes were operated both with gas flowing and at rest. The two oxides of nitrogen exhibited strongly the third positive group of nitrogen bands especially when the gas was flowing. This group of bands was observed from 190.2 to 345.8mjj,. Of the two negative groups of carbon bands the first appeared probably due to carbon monoxide and the second to carbon dioxide. Bair discovered several new bands in this second group and forty new bands in the spectrum of sulphur dioxide, extending this group to 212.4m|i. 160 ULTRAVIOLET RADIATION Hoist and Oosterhuis 81 have described experiments which appear to show that the so-called cyanogen bands are not due to nitrogen but to one of its compounds which condenses at a much higher temperature, probably cyanogen. In some of the experiments the discharge tube was immersed in liquid oxygen and the spectrogram was obtained through the walls of the Dewar vessel. On one spectrogram the bands 385.5, 388.3, and 416.8mpi appeared but the others were absent. It thus appears possible that the bands are due to two different carriers. The origin of the cyanogen bands has been investigated by Barratt 82 by observing the flame spectra of a number of gases containing carbon, hydrogen, nitrogen and oxygen. The cyanogen bands are strongly developed in flames of coal-gas and nitrous oxide, of coal-gas and air, of carbon monoxide, air and ammonia, of HCN and air, of methylamine and air, and other nitrogenous organic substances. The bands are absent from the flame of hydrogen and nitrous oxide if all traces of carbon are excluded and are not found in hydrocarbon-oxygen flames in general or in ammonia-oxygen flame. Carbon is es- sential to the production of the bands and the appearance of the cyanogen bands is a delicate test for carbon and for compounds of nitrogen admitted in the form of a gas to hydrocarbon flames burning in air. The intensity of the cyanogen bands when carbon compounds are ad- mitted to the hydrogen-nitrous oxide flame was not found to bear a simple relation to the amount of carbon added. Anderson 83 has described a method of obtaining high temperatures for laboratory purposes which may have some applications in the ultraviolet region. The method consists in electrically exploding a fine wire in a con- fined space. When the explosion occurs in air confined in a tube or slot the flash gives a brilliant continuous spectrum crossed by the absorption lines of the elements of which the wire is composed. Iron, copper, nickel, EXPERIMENTAL SOURCES 161 and manganin so far have been investigated. It is hoped that it may be possible by this means to imitate stellar absorption spectra of the solar type. By discharging a large condenser, charged to 26000 volts, through a fine wire 5 cm. long, about 30 calories of energy were dissi- pated in about one hundred thousandth of a second. If all this energy had entered the two mgm. of wire it would have raised its temperature to 300000 C. According to Anderson the brilliant flash possessed a brightness corresponding to a temperature of about 20000 C. or about 100 times the brightness of the sun. A method of producing fine wires on a lathe is also described. The spectrum of the iron arc has been recently investi- gated by H. Schumacher 84 who used a vertical arc 6 mm. in length and a current of 4 amperes at 20 volts. He checked his results with the arc-lines of copper, silver, and nickel and compared his observations with those of Kayser and Runge. St. John and Babcock 85 have de- termined the wave-lengths of the lines of the iron arc by means of grating and interferometer measurements be- tween 337 and 675m[A. They have presented a table of 1076 lines most of which were measured on many spec- trograms. Strutt 8G used a sodium-vapor arc in quartz in studying the line-spectrum of sodium as excited by fluorescence. He found that the excitation of sodium vapor by the second line of the principal series leads to the emission of line, 330.3mji, and the D line. He also studied the absorption of the vapor. Polarization could not be de- tected in the ultraviolet resonance radiation, though it has been readily observed in D resonance radiation. Datta 87 used a similar lamp in a study of the vacuum arc-spectra of sodium and potassium. 162 ULTRAVIOLET RADIATION References z. Handbuch d. Spectroscopie. 2. Atlas of Emission Spectra, 1905. 3. Spectroscopy of the Extreme Ultraviolet, 1914. 4. Smithsonian Physical Tables. 5. Color and Its Applications, 1921. 6. J. Opt. Soc. Amer. 4, 1920, 496. 7. Phys. Rev. i, 1913, 329. 8. Astrophys. Jour. 33, 191 1, 98. 9. Phil. Mag. 41, 1921, 814. 10. French patent 419117. 11. Comp. Rend. 158, 1337. 12. Lancet 1917, 996. 13. Phys. Rev. 12, 1918, 167. 14. Sitz. Heid. Akad. Wiss. 1910. 15. Trans. Roy. Soc. London, 214, 1914, i. 16. Phys. Rev. 8, 1916, 674. 17. Ber. Akad. Wis. Wien. 102, Ila, 438 and 694. 18. Inaug. Dis. Berlin, 1909. 19. Astrophys. Jour. 35, 1912, 341. 20. French patent 468215. 21. Z. Wiss. Phot. 2, 1904, 31. 22. Ann. d. Phys. 13, 1904, 901. 23. Astrophys. Jour. 38, 1913, 282. 24. J. Amer. Chem. Soc. 1906, 786. 25. Met. and Chem. Eng. 18, 1918, 232. 26. Trans. Amer. Electrochem. Soc. 1917. 27. Phil. Trans. 202, 1904, 430. 28. J. Frank. Inst. 185, 1918, 552. 29. Chem. Abs. 1917, 235. 30. Chem. Engr. vol. 26. 31. U. S. patent 1110576. 32. U. S. patents 1091244 and 1110574. 33. Phys. Rev. 30, 1910, 641. 34. Comp. Rend. 155, 1912, 852. 35. J. Biol. Chem. 20, 1915, 315. 36. Ann. d. Phys. 10, 1903, 446. EXPERIMENTAL SOURCES 163 37. Elec. Rev. 67, 1916, 654. 38. U. S. patent 1197629. 39. Elec. Rev. 66, 1915, 1055. 40. Proc. Amer. Acad. Arts and Sci. 51, 1916, 637. 41. Comp. Rend. 156, 1063. 42. Trans. Chem. Soc. 107, 1915, 682. 43. Phys. Zeit. 1910, 1039. 44. Z. Phys. Chem. 80, 235. 45. Astrophys. Jour. 52, 1920, 47. 46. Astrophys. Jour. 53, 1921, 150. 47. Comp. Rend. 171, 1920, 1106. 48. Comp. Rend. 170, 1920, 31. 49. N. Cimento, 18, 1919, 82. 50. Roy. Soc. Proc. 98, 1920, 114. 51. Roy. Soc. Proc. 95, 1919, 316. 52. Roy. Soc. Proc. 98, 1920, 95. 53. Proc. Roy. Soc. 98, 1920, 109. 54. Roy. Soc. Proc. 95, 1919, 258. 55. Bull. Soc. Chem. 25, 1919, 305. 56. Astrophys. Jour. 49, 1919, 224. 57. Comp. Rend. 171, 1920, 320, 709 and 909. 58. Comp. Rend. 172, 1921, 803 and 851. 59. Zeits. Wiss. Phot. 19, 1920, 289. 60. Zeits. Wiss. Phot. 19, 1919, 149. 61. Phil. Mag. 39, 1920, 457. 62. Akad. Wiss. Wien Ber. 127, 1918, 1099. 63. Chem. Zentralbl. 1917, 362. 64. Zeits. Wiss. Photochem. 16, 1917, 157. 65. Zeits. Wiss. Photochem. 17, 1917, 132 and 1918, 145. 66. Roy. Soc. Proc. 95, 1919, 438. 67. Zeits. Wiss. Photochem. 18, 1918, 45. 68. Zeits. Wiss. Photochem. 18, 1918, 84. 69. Zeits. Wiss. Phot. 19, 1919, 129 and 142. 70. Phys. Zeits. 21, 1920, 25. 71. Bull. Bur. Stds. 14, 1918, 765. 72. Ann. d. Physik. 60, 1919, 405. 73. Ann. d. Physik. 58, 1919, 333. 74. Ann. d. Physik. 63, 1920, 201. 75. Phys. Zeit. 21, 1920, 638. 164 ULTRAVIOLET RADIATION 76. Roy. Soc. Phil. Trans. 220, 1919, 137. 77. Nature, 104, 1919, 314. 78. Nature, 104, 1919, 393. 79. Astrophys. Jour. 52, 1920, i. 80. Astrophys. Jour. 52, 1920, 301. 81. K. Akad. Amsterdam, Proc. 23, 1921, 727. 82. Roy. Soc. Proc. 98, 1920, 40. 83. Astrophys. J. 51, 1920, 37. 84. Zeits. Wiss. Phot. 19, 1919, 149. 85. Astrophys. J. 53, 1921, 260. 86. Roy. Soc. Proc. 96, 1919, 272. 87. Roy. Soc. Proc. 99, 1919, 69. Relative. Quartz Mercury Arc Exposure Cobalt Glass I n a Z n n 4 Clear Glass 4 n n 2 n n 1 Quartz Mercury Arc Iron Arc 1 Clear Glass Cobalt " }' Clear n Cobalt " }' Clear Cobalt " }< Clear " Cobalt n }> Clear Cobalt a }' Quartz Mercury Arc Plate VIII. Ultraviolet transmission spectra of clear and cobalt glasses as obtained by a quartz prism spectrograph. The sources of radiation were the quartz mercury arc and the iron arc. Both glasses were of the same composition with the exception of the addition of a slight amount of cobalt to one of them. CHAPTER X DETECTION AND MEASUREMENT A great variety of means is available for detecting ultraviolet radiation and many of these may be utilized for measurements. The absolute measurement of radia- tion can be directly achieved by means of instruments such as the bolometer, thermocouple, thermopile, and radiometer, which measure incident radiation. Absolute measurements can be obtained also by indirect comparison methods in which the photographic plate, the photoelectric cell, the phenomenon of phosphorescence, the selenium cell, and a large number of photo-chemical reactions may be utilized. The types of apparatus and the methods for measuring ultraviolet radiation are restricted by the characteristics of this radiation. One of the greatest re- strictions is the relatively smaller quantities of energy usually encountered than in investigations of visible and infra-red radiations. On the other hand, there are many effects peculiar to ultraviolet radiation which do not attend infra-red radiation. The same comparison can be drawn between ultraviolet and visible radiation although these two spectral regions have much in common. In general, any effect which is produced by ultraviolet radiation may be utilized in obtaining measurements per- taining to the latter. These effects may be photo- chemical, photogenic, physical, physiological, germicidal, photo-electrical, etc., although there is more or less over- lapping of these in many phenomena. In fact, these di- visions themselves are not strictly independent of each other. It is not the intention to discuss the methods of measurement in detail because previous chapters con- tain much pertinent data and detailed accounts of the 165 ' 166 ULTRAVIOLET RADIATION various instruments may be found elsewhere. The data in other chapters pertaining to reflectivity and trans- parency of various media indicate the limitations and uses of these media in the measurement of ultraviolet radia- tion. The determination of the spectral characteristics of ultraviolet radiation is one of the most essential kinds of data upon which to base conclusions and to depend for future progress, but by dispersing radiation into its spec- trum, the ability of energy-measuring instruments to record the enfeebled radiation is very seriously taxed. In many cases it is quite sufficient to use filters such as quartz, glass, etc., and to measure total radiation, provided the source and its conditions of operation are accurately de- scribed. In such cases the thermopile, the bolometer, and the radiomicrometer, are usually sufficiently sensitive. Scientific literature abounds with references to the de- velopment and use of such instruments and their acces- sories but it is not a function of this book to discuss these works. Coblentz x has presented an excellent discussion of these instruments which includes many details of con- struction and operation. Nutting 2 has treated these, the photographic plate, and radiation laws. Langley developed the bolometer which consists essen- tially of a blackened strip of metal. This strip absorbs the radiant energy and its temperature is therefore increased. The temperature-rise is determined by noting the change in electrical resistance. This strip has been placed in a vacuum with a consequent improvement in operation. Coblentz advises a vacuum of not more than 0.1 mm. mercury pressure. Among the many contributions on vacuum bolometers are those by Paalzow and Rubens, 3 Buchwald 4 and Warburg, Leithauser, and Johansen. 5 Coblentz described the difficulties due to drift caused by unequal warming of the strips and to the variations caused by air-currents. He compared the bolometer with the thermopile. DETECTION AND MEASUREMENT 167 The name of Rubens is associated with the development of the thermopile perhaps more than any other. The thermopile consists of thermocouples arranged so that the effect of a single couple is augmented by that of others. It consists essentially of a number of thermoelements or thermocouples connected in such a manner that the proper ones are heated by the incident radiation. The difference in temperature between these junctions and the cooler ones causes a current to flow. A sensitive galvanometer records the current which is a measure of the incident radiant energy. The thermopile is used either in a vacuum or in a screened chamber open to the air. Iron-constantin couples have been extensively used but copper-constantin, bismuth-silver and bismuth-iron are also employed. The thermoelectric power of copper-constantin is somewhat lower than that of iron-constantin but the copper pos- sesses the advantage of not rusting. Coblentz * has pre- sented valuable details, pertaining to the various kinds of thermopiles, drawn from his extensive experience. He found that the Rubens thermopile was only about one-half as sensitive as a bolometer but that it could be improved by using thinner wires and by placing it in a vacuum. The thermopile is more sluggish than the bolometer and therefore is less adapted to work requiring instan- taneous registration. However, on account of its greater steadiness it is recommended for measuring feeble radia- tion such as in the ultraviolet regions of the spectrum. The two instruments are about equally efficient in meas- uring radiant energy; that is, they are about equally sen- sitive to radiant energy. Pfliiger 6 employed the thermopile successfully in the measurement of ultraviolet radiation as far as 186mji. The radiometer devised by Crookes 7 is familiar to many as an interesting toy, but even in the original form of rotating vanes it can be used to measure radiation of 168 ULTRAVIOLET RADIATION sufficient intensity at least approximately. A relation between revolutions per second and intensity of radiation can be obtained. Crookes fastened pieces of pith, one black and the other white, at the ends of a long straw. This was suspended by a silk fiber in a glass tube. The later ones seen in optical shops consist of upright vanes of mica blackened on one side. This " paddle wheel " is suspended upon a pin by a small cup of glass so that fric- tion is reduced to a minimum. The glass bulb is ex- hausted to a low pressure. The rotation is caused by the unequal bombardment of the two sides of the vone by the molecules of the rarefied gas. The blackened side be- comes warmer than the other by the absorption of heat and the gas molecules in contact with the blackened side have imparted to them greater kinetic energy than those on the other side. The net pressure of the bombardments causes the wheel to rotate with the blackened sides of the vanes hindmost. Although the radiometer described in the preceding paragraph has been used to measure radiant energy, Nichols 8 was the first to make a sensitive instrument em- ploying this principle. He employed two blackened vanes of mica or of thin platinum fastened to a horizontal arm. This was suspended by means of a very fine quartz fiber. Radiation is permitted to fall upon one of the vanes and this causes a tendency to rotate the arm. A mirror at- tached to the suspended system reflects the image of a scale and in this manner the deflection is measured. The sensitiveness of this instrument is a function of the pres- sure of the gas, of the kind of gas, and of the distance of the vanes from the window of the enclosure. The pres- sures employed are a few hundredths of a millimeter of mercury. Coblentz 9 has discussed this instrument in detail. According to him, the instrument is not selective and is as efficient in the ultraviolet as is the bolometer. In this same paper Coblentz also presented many details pertaining to the bolometer. DETECTION AND MEASUREMENT 169 The radiomicrometer was developed by Boys 10 and by d'Arsonval. 11 The former used a loop of copper wire to which a junction of bismuth-antimony was soldered. The latter used a loop, one-half of which was silver and the other half was palladium. The instrument is essentially a moving-coil galvanometer having a single loop of wire with a thermojunction at one end. Fery 12 made what he termed a radiomicrometer which consisted of a loop of copper joined at the bottom by means of a piece of constantin wire. The two junctions were placed at the same height, side by side, and covered with thin strips of silver. The latter were polished on one side and blackened on the other. The deflection of this sys- tem, properly screened, is a measure of the incident radia- tion. Schmidt 13 made such an instrument with a sus- pended system consisting of bismuth-antimony. Cob- lentz 14 has devised various improvements. He found that by placing it in a vacuum its sensitiveness was almost double. He showed that the highest efficiency is ob- tained when the resistance of the thermocouple is equal to the combined resistance of the connecting wires and of the auxiliary galvanometer. According to Coblentz 9 the sensitiveness of the galvanometer is very limited and is perhaps only one-fifth that of the best bolometers. It is not a promising energy-measuring instrument for the ultraviolet. This leaves the field chiefly to the bo- lometer and the thermopile. The auxiliary galvanometer which is necessary in the use of the bolometer, the thermopile, and the radiomi- crometer, is one of the difficult obstacles. Various in- vestigators have given much attention to its develop- ment. Valuable discussions will be found in the works of Coblentz cited in the preceding paragraphs. Of course, a thermometer with a blackened bulb can be used to measure radiation but it is not sensitive enough to be considered in the same company with the 170 ULTRAVIOLET RADIATION thermopile and the bolometer. However, Callender 15 recently described a thermoelectric balance for the meas- urement of radiation. A thermojunction in the form of a disk is exposed to radiation and the rise in tempera- ture is opposed by the well-known cooling (Peltier) ef- fect produced by sending an electric current through the junction. Another thermocouple in contact with the disk indicates the amount of compensation. This instru- ment has been termed a radiobalance. For details per- taining to its construction the original paper should be consulted. Coblentz 1 has also discussed it at length. Weber 16 has described a micro-radiometer which forms two arms of a Wheatstone bridge. These arms consist of a narrow glass tube containing a drop of mercury at the center and solutions of zinc sulphate at the ends, into which platinum electrodes are immersed. The ends of glass tube are large bulbs containing air and having rock-salt windows. One of the bulbs is coated inside with lamp-black or platinum black, with the exception of the window, and outside with an opaque non-conduct- ing material. When radiation is permitted to enter the window of the blackened bulb the air is expanded and the liquids are pushed toward the other bulb. This al- ters the relative lengths of the column of mercury and of the zinc sulphate solutions between the platinum ter- minals which unbalances the bridge owing to the change in the resistances of these two arms. The instrument does not appear to be sensitive enough for the measure- ment of ultraviolet radiation spectrally, but it may have some use in the measurement of total ultraviolet or other radiation of sufficient intensity. It is well known that the resistance of selenium changes when exposed to radiation and this phenomenon has been utilized in the measurement of radiation. How- ever, selenium is very selective in this respect, it being most sensitive to long-wave visible radiation and much DETECTION AND MEASUREMENT 171 less sensitive to ultraviolet radiation. A photoelectric cell, on the other hand, is usually more sensitive to the energy of shorter wave-lengths than to that of long- wave visible radiation. 17 The photoelectric cell is also selective in its action. In the case of the blackened parts of the energy-measuring instruments discussed in the preceding paragraphs, there is practically no selectivity. The black coat, which is commonly lamp-black or plati- num-black, is non-selective, at least in the near ultraviolet, visible, and infra-red regions. Its chief fault is that it does not absorb all the incident radiation ; that is, it is not perfectly black. This error is usually very small and the procedure can be such as to eliminate its effect. Some of the various forms of energy-measuring instru- ments already described are quite sensitive enough to measure total ultraviolet radiation between certain spec- tral limits, but only the thermopile and bolometer appear to have wide applications in spectral energy measure- ments. Even these must be given every advantage of experience, skill in manipulation, and sensitiveness of gal- vanometer. Pfliiger 18 obtained spectral energy measure- ments of the radiations from arcs and sparks by means of a bolometer and a fluorspar prism. He was able to measure the relative energy of the strongest lines as far as 186mpi but he was unable to measure the fainter ones. He 19 also used the thermocouple in determining the spec- tral absorption-factors of various substances. The great advantage of measuring energy directly has led various investigators to push the energy-measurement instrument to the limit of its ability. Hagen and Rubens 20 employed the thermopile as far as 250mpi in their studies of the spectral reflection-factors of metals. Other references to such applications are found in other chapters. As previously stated, the effects of radiation between certain wave-lengths can be studied by means of filters 172 ULTRAVIOLET RADIATION and instruments which measure total radiant energy. This suffices in many cases and certainly is much better than making no attempt at isolation. The ideal in many other cases is to make refined spectral investigations by means of the spectroscope. The reflection grating pos- sesses the advantage of producing a normal spectrum. For absolute measurements the selectivity of the sub- stance upon which the grating is ruled must be known. The concave grating is readily applicable to the near and middle ultraviolet regions, that is, for radiation as short as 200mj4, in wave-length. Air is opaque to the extreme ultraviolet so that Schumann, the pioneer in- vestigator in this region, constructed a vacuum grating spectrocsope. 21 Lyman 22 constructed a vacuum grating spectroscope and was the first to make successful measurements of wave-lengths in the extreme ultraviolet region. His work has been very extensive in this region and much of the data available is due to his ingenuity and persist- ency. The grating spectroscope has been a popular instru- ment for invading unexplored regions because of its normal spectrum. When a prism is used, its dispersion curve must be known. This involves much uncertainty in unexplored regions. Morris- Air ey 23 was one of the first to devise a transmission-grating for the ultraviolet. He ruled a plate of fluorite for this purpose but his in- vestigation was not very successful. Others have tried the fluorite grating with mediocre success. The concave reflection grating in a vacuum has been the most success- ful apparatus for the extreme ultraviolet. McLennan and Lang 2 * among others have investigated the extreme ultraviolet with a vacuum grating spectro- scope. Millikan 78 and his colleagues 79 were able to extend the study of the spectra of certain elements far into the DETECTION AND MEASUREMENT 173 extreme ultraviolet chiefly by the perfection of a vacuum spectrometer and of specially ruled gratings. A mercury diffusion-pump was connected with the spectrometer enclosure and was operated continuously notwithstand- ing the fact that the source of radiation, which was a very high potential spark, was operated intermittently. The spark-gap varied in length from 0.1 mm. to 2 mm. and the sparking was accomplished in the high vacuum by means of a battery of Leyden jars charged to a poten- tial of several hundred thousand volts obtained from a powerful induction coil. The gratings were ruled with great precision and with a light touch so that about half the original surface was left between the rulings. It was not found practical to increase the number of lines to more than 1100 per millimeter. In the ordinary process of ruling gratings for work in the visible spec- trum the surface of the grating is entirely cut away by the ruling-diamond with the result that most of the radi- ation is thrown into spectra of higher order than the first. For work in the extreme ultraviolet the overlap- ping of spectra renders all but the first order almost use- less, so that it is very desirable to produce a first-order spectrum as intense as possible. Results obtained with this apparatus are presented in Chapter IX. Merton 80 has described a method of spectrophotometry applicable to any part of the spectrum which can be photographed through quartz lenses and prisms. This method consists in crossing the dispersing system with a very coarse grating and reducing the length of the slit to a very small value. The grating is placed between the prism and the camera lens of the spectrograph with the lines of the grating perpendicular to the refracting edge of the prism. As a result of this arrangement a continuous spectrum appears on the photographic plate as a dark central strip with a succession of other strips of different intensities on either side. The intensities of 174 ULTRAVIOLET RADIATION these orders are determined by the ruling of the grating, and the width of the strips by the length of the slit. In the case of a " line " spectrum the spectral lines are re- corded on the plate as dots of different densities on both sides of the central dot. If the last dots which are just visible in the case of two lines are noted, a previous knowledge of the relative intensities of the different orders corresponding to these dots makes it possible to determine the relative intensities of the lines. Owing to the fact that the slit is very small (approaching a point in size), different regions of a light-source may be in- vestigated and for the same reason the method may be applied to the study of stellar spectra. Merton made and examined a number of gratings and found the most convenient ruling for use with the quartz spectrograph to be about 25 lines to the inch. Wire- wound gratings are not feasible in this case because a very small rotation of the grating about an axis parallel to the rulings considerably alters the distribution of in- tensity in the different orders by changing the ratio of the transparent to the opaque parts of the grating. Gratings employed for this purpose should not be made by cutting grooves in a transparent surface owing to the characteristics introduced by irregularities and another disturbing factor, not independent of wave-length, which is due to the form of the groove. Merton coated a quartz plate with a very thin layer of lamp-black by holding the plate over burning toluene and then flowed over the plate a thin mixture of alcohol and shellac. The ruling was done by means of a bone tool in a shaping machine possessing an automatic feed. Lord Rayleigh's formula 81 can be applied to gratings consisting of alter- nate transparent and opaque bars for computing the brightness of any order. Merton calibrated his gratings by means of a neutral wedge for which the density-step as a function of wave-length had been determined as de- DETECTION AND MEASUREMENT 175 scribed in a previous investigation. 82 As a source of light for this calibration the mercury blue line, 435.9mji was used. L. and E. Bloch 83 have investigated the spectra of many elements by means of a vacuum spectrograph with lenses and prism of fluorite. The apparatus is enclosed in a brass casting 2 cm. thick and is closed by three thick brass plates disposed before the prism, slit, and photographic plate respectively. The plate opposite the slit contains a fluorite window so that the source of radiation can be located outside the chamber. The latter is exhausted to a pressure less than 0.001 mm. With this apparatus spark-spectra of different metals have been investigated down to about 140m(i. They employed a condensed spark in hydrogen at atmospheric pressure. The quartz spectrograph is perhaps the most generally useful apparatus for work with the near and middle ultra- violet. Its transparency extends to about 185mpi. Several types are available but it is easy to make one if the essen- tial optical parts, a prism and two lenses, are available. In fact, sometimes it is advantageous to use two prisms. Hilger supplies a small spectrograph which is very useful and a large one which yields an ultraviolet spectrum about 18 cm. long. In all these prism instruments the spectrum is brought to a focus in a plane considerably inclined to the optical axis owing to the chromatic aberration of the simple lenses. This obliquity can be utilized to advan- tage in some investigations by cutting a window in the side near the acute angle. Through this, fluorescent spectra of solids, for example, may be viewed and even photographed. By projecting the spectrum downward at the proper angle it may be brought to focus on the surface of fluorescent liquids and the latter may be studied in this manner. For this purpose one or two prisms and two lenses of quartz sometimes can be used more advan- tageously than a complete instrument. 176 ULTRAVIOLET RADIATION Houston * 5 has described a spectroscope suitable for work in the ultraviolet, visible, and infra-red regions al- though no new fundamental principle is involved. Radi- ation from a slit is rendered parallel by a concave mirror of nickel. It then passes through a prism of quartz, glass, or rock-salt, depending upon the region to be studied, and the spectrum thus formed is brought to a focus by another concave nickel mirror which may be rotated. According to Houston, the nickel mirrors are only about one-half as efficient as quartz lenses, but owing to better collimation, larger apertures may be used than in the case of lenses. There are certain advantages in the focusing of the mirrors. A rhomb of quartz is used to divide the beam of radiation and surfaces of rough- ened silica diffusely reflect the energy. The two spectra are juxtaposed on the photographic plate or in the eye- piece. Lankshear 26 has described an instrument consisting of two identical quartz systems of reflecting prisms and lenses which focus two beams of radiation from the same source upon the slit of a quartz spectrograph. The medium to be studied for ultraviolet absorption is placed in the path of one of the beams and a novel sector is placed in the path of the other. The sector is designed to avoid intermittent illumination which sometimes casts doubt on photographic results. Fluorescent substances are very convenient for focus- ing or for examining ultraviolet spectra. A fluorescent uranium plate glass is one of the best for this purpose. The phenomenon of fluorescence has been utilized con- siderably for visual observations in the ultraviolet. It performs a function similar to that of the photographic plate. It makes it possible to observe the ultraviolet visually which is quite desirable in many cases. In this manner approximate transparency limits of substances can be determined quickly if a powerful source of ultra- DETECTION AND MEASUREMENT 177 violet radiation is focused upon the slit of the spectro- graph by means of a quartz lens. It does not possess the advantage of permanent record as the photographic plate does. Various applications of fluorescence have been made in instruments for the study of ultraviolet radiation. The calibration of spectroscopes for the ultraviolet is now a comparatively simple matter because there are many well-determined spectral lines available for that purpose. The grating spectroscope presents no difficulty in this respect because of the normal spectrum. It is only necessary to photograph a pure spectrum of a spark, a discharge tube, or an arc whose spectral lines have been carefully determined by some one. The data pertaining to spectra are found in various references which have been given in preceding chapters. When it is necessary to know the refractive-indices of various substances such as quartz and fluorite, these can be found in the various optical tables already cited. One of these 27 which is quite accessible gives the refractive- indices of fluorite, Iceland spar, rock-salt, sylvine (potas- sium chloride), and quartz for various wave-lengths as far into the ultraviolet as lS5m\i. Stark 28 determined the refractive-indices of fluorite from 185 to 656mpi. Martens 29 obtained the refractive- indices of the ordinary and the extraordinary rays from 198 to 768mpi, and the refractive-indices of rock-salt and of sylvine from 185 to SOOmji. Handke 30 determined the indices of refraction of fluorite as far as ISlmpi. Owing to the importance of quartz and the prevalence of quartz prisms in ultraviolet investigations the refrac- tive-indices as determined by various investigators are presented in Table XXXIV. These are mean values 27 in air at a temperature of 18 C. for the ordinary and the extraordinary rays. In the table n and n e are the refrac- tive indices for the ordinary and the extraordinary rays respectively. 178 ULTRAVIOLET RADIATION TABLE XXXIV Index of Refraction of Quartz m/i n n e 185 1.67682 1.68999 193 .65997 .67343 198 .65090 .66397 206 .64038 .66300 214 .63041 .64264 219 .62494 .63698 231 .61399 .62560 267 .59622 .60712 274 .58762 .59811 340 .56748 .57738 396 .55815 .66771 410 .56650 .56600 486 .54968 .55896 589 .54424 .55334 656 .54189 .55091 686 .54099 .54998 760 .63917 .54811 Disch 31 was one of the first to employ the mercury lamp in polarimetry. The green line of the mercury spec- trum is exceptionally pure, and where a high intensity of monochromatic radiation is desired this line is easily sepa- rated by means of filters which are less wasteful than the spectroscope in isolating monochromatic radiation. In polarimetry, very high intensity is not usually required, so that Lowry 32 has used a globule of mercury in a hydro- gen discharge tube. Such a tube starts readily and by heating the mercury, its spectrum can be increased in brightness if necessary. Sirks 33 has determined the rotation of the plane of polarization for ultraviolet radiation of various wave- lengths for hydrogen, oxygen, and carbon dioxide. His determinations extended as far as 238mpi. DETECTION AND MEASUREMENT 179 Photography is one of the most helpful allies in the study of ultraviolet radiation. There are many emulsions avail- able on glass plates, celluloid films, and paper, and the choice will be influenced by the work to be done in a particular case. In general, ordinary emulsions are satisfactory for the photography of radiation of wave- lengths between 200 and SOOmpi. In the case of the photographic emulsion it is necessary to establish rela- tions between the density of the photographic image, the time exposure, and the intensity of radiation. This is necessary for each wave-length for accurate work. Schwarzschild's law is expressed thus, It p = constant where I is the intensity of the radiation, t is the time of exposure, and p is an exponent whose value lies usually between 0.7 and 1.0 for various emulsions. For example, if the intensity of radiation in one exposure is only one- half the value for another exposure, the former exposure must be slightly greater than twice that of the other in order to produce the same photographic density. After plates have been developed they are measured for transparency and the density is established by the following relation, D = log O = log ^ where D is the density, O is the opacity, and T is the transmission-factor. The curve obtained by plotting density and the logarithm of the intensity of radiation has a straight portion which is considered to be the best region of exposures or densities. This straight-line rela- tion was discovered by Hurter and Driffield in their pio- neer researches. The range of this straight portion of the curve for any given case is a measure of the latitude of the emulsion under the conditions of development. 180 ULTRAVIOLET RADIATION Sheppard and Mees, 3 * Nutting, 2 Sheppard 35 and others have treated the photographic process so thoroughly as to obviate the necessity of doing so here. In general, in the photography of the ultraviolet it is best to produce images of the same photographic density on the same plate and by equal exposures. One of these images is preferably produced by radiation of known intensity, although in many cases relative inten- sities are satisfactory. On producing equal densities with equal exposures it is obvious that relative (or abso- lute) intensities of radiation are obtained directly. Usually the inverse-square law is quite dependable if the distance from the source (a diffusing medium near the slit of the instrument), or from the photographic plate (when no spectroscope is used) is at least ten times the largest projected dimension of the source. In other words, if the source is a mercury tube two feet long, the distance should be at least 15 feet and preferably 20 feet if the inverse-square law is to be employed. Of course, when sufficient density is obtainable it is best to screen most of such a source, thereby obtaining an effective source of small dimension. Care must be taken to avoid attempting to use the inverse-square law when bright- ness of the source instead of illumination due to it is employed. It is difficult to obtain non-selective screens for re- ducing the intensity. When this is necessary, however, a wire mesh or several of them can be employed in some cases with success. The sectored disk has been widely employed but doubt may arise as to the relation of the sector openings to the photographic action. It is well to test this in any case, and even better to eliminate this source of doubt by employing the same sector opening if possible for any comparison. Care must be taken to provide fog-strips when necessary in order to be able to subtract the density due to unavoidable fog, some of which is inherent in the emulsion. DETECTION AND MEASUREMENT 181 Photographic attachments can be purchased for spec- troscopes. These include not only the camera but sec- tored disks with various openings. The Hilger sector- photometer is quite satisfactory. Howe, 36 among others, has described investigations using the sector-photometer in combination with the quartz spectrograph. Tyndall 37 has described some minor improvements. The neutral wedge is quite useful. If placed before the slit of a spectroscope so that its transparency varies along the length of the slit, the spectrogram will reveal very roughly the spectral characteristic. That is, the spectro- gram will vary in height throughout its length depending upon the intensity of radiation and photographic sen- sibility. A non-selective wedge can be made by flowing a gela- tine solution of a neutral dye upon an inclined plate of quartz or glass. Nigrosine is sometimes nearly non- selective for visible and near ultraviolet radiation. The measurement of photographic density can be ac- complished by various photometric methods. Owing to the smallness of the areas, the optical pyrometer affords a convenient method. If the areas involved are large enough, an ordinary photometer can be used; that is, relative brightnesses are measured. The Martens polarization photometer is a very convenient device for this purpose. Fabry and Buisson 38 among others have described a microphotometer convenient for determining photo- graphic densities. They compare the transparency of the image with that of various portions of a thin wedge of neutral glass. Henri and Wurmser 39 described spectrophotometric measurements by the method of equal densities. They varied the time of exposure until the density of the image produced by the direct beam of radiation was equal to that resulting from the radiation which has been absorbed 182 ULTRAVIOLET RADIATION to some extent. By using Schwarzschild's law a sufficient degree of accuracy is obtained for most purposes. Nutting 40 has described a method of photographic spec- trophotometry in which radiation from each of two beams produces a series of interference bands. The bright bands of one series are superposed on the dark bands of the other series and when the intensities are equal the bands disappear. The tungsten lamp is very useful for investigations in the near ultraviolet when a continuous spectrum is de- sired. By using a quartz bulb or window its usefulness extends into the middle ultraviolet. Some details per- taining to modern tungsten lamps and various plates are to be found elsewhere. 41 Difregger 42 has developed a method of decreasing the intensity of radiation in a continuous manner and at the same time moving the photographic plate so that it re- ceives a correspondingly decreased intensity. The author 43 a number of years ago devised a method for approaching the ideal result obtainable with an uni- form energy spectrum and a photographic emulsion of uniform spectral sensibility. By photographing a con- tinuous-spectrum and developing the image to a proper density, this spectrogram may be placed in its correct position in the plate-holder so that spectra thereafter are photographed through it. This automatically compen- sates approximately for non-uniform spectral distribution of energy and non-uniform spectral sensitiveness of the emulsion. It is especially useful in obtaining the absorp- tion spectra of substances. The " spectrophotographic filter " is made by placing the plate in the holder with the emulsion away from the prism or slit so that when it is replaced in the plate-holder the spectrogram is in contact with the emulsion of the plates upon which absorption- spectra are to be photographed. In the original paper various results with prism and grating spectrographs are DETECTION AND MEASUREMENT 183 shown. Of course such a " filter " is useful only for the region for which glass is transparent. For use in the middle ultraviolet it would be necessary to make it on a quartz photographic plate. For work in the near and middle ultraviolet regions, ordinary emulsions are usually satisfactory; however, for ultraviolet radiation of shortest wave-lengths, gelatine is opaque. Schumann 44 has described a method for making dry plates for the extreme region but it is not necessary at the present time for the investigator to make his emul- sions unless he is concerned with the remote portion of the extreme ultraviolet region. Plates should be of fine grain and free from fog and defects. The spectral limits of sensitiveness of photographic plates and films are due to the opacity of the gelatine and not to the failure of the silver salt to respond. Schumann used a specially prepared emulsion of silver bromide very weak in gelatine. These plates were not sensitive to radiation longer than SOOmjx in wave-length. Such emulsions are the only means at present for studying the ultraviolet region of shortest wave-lengths. The speeds of commercial photographic plates and papers vary over a wide range. Plates of extremely high speed are more than thirty times faster than the slowest plates. The plate commonly used for lantern slides is of fine grain but more contrasty than the more common plate. Fast bromide paper is a thousand times faster than the slowest commercial paper. Usually the slow emulsions are of fine grain which in- fluences the resolving power. The latter is also greatly affected by the developer. Pyro and hydroquinone are among the best developers for obtaining high resolving power. Huse 45 has published results obtained with present-day plates and developers. Using pyro-soda developer he found for several plates, the number of lines per millimeter which is just resolvable. The lines were 184 ULTRAVIOLET RADIATION opaque and separated by spaces of the same width. His results for typical plates are: albumen 125, resolution 81, process 67, lantern 62, medium speed 35, high speed 27. Using a lantern-slide plate he found the resolving power of developers to vary from 77 to 47. In Chapter IV references are made to atlases of absorp- tion media. Many data pertaining to filters will be found elsewhere 17 but there is not a great deal available for the ultraviolet. Various preceding chapters contain much of interest pertaining to filters which can be employed in photography. Recently Hodgman 46 has presented vari- ous data pertaining to gelatine filters between glass plates cemented together with balsam. The flowing solution consisted of six per cent by weight of clarified gelatine, a definite quantity of one or more aqueous dye-solutions, and distilled water. The data which he presents includes the dye used, the strength of the aqueous dye-solution, the quantity of this solution used in the final mixture, the quantity flowed per unit area, the wave-length limits of action of light upon various types of photographic plates. The panchromatic plate used was acted upon by the radia- tion from a " Mazda C-2 " lamp after passing through glass, between 350mjx and 720mpi. A region of low sensitivity existed for this plate in the vicinity of 520mji. The ordi- nary plate used exhibited a range of action, under the con- ditions of the investigation, between 350mpi and 550m^i. The action on the orthochromatic plates extended to 63Qm\ji. The data pertaining to 41 filters are presented and they include red, orange, yellow, green, blue, pink, violet and purple filters. Of course, any photo-chemical reaction can be used for measuring the intensity of the radiation involved in pro- ducing the reaction. Many of these are discussed in other chapters but a few others may be of interest. Ultraviolet radiation causes discoloration of filter-paper moistened with a 20 per cent solution of potassium ferro- cyanide. DETECTION AND MEASUREMENT 185 According to Schall, 47 paper prepared with p-phenylene- diamine nitrate (f normal) is sensitive only to radiation shorter than 313mji in wave-length. Inasmuch as the solar spectrum does not extend beyond 290m^i, this paper may be used for studying the variation in the ultraviolet energy of the shortest wave-lengths present in solar radia- tion. The intensity of violet and ultraviolet radiation can be measured by the rate of decomposition of oxalic acid in the presence of uranyl acetate. According to Freer and Gibbs 48 this reaction has a very small temperature coef- ficient. The effect of sunlight in promoting the color- ation of benzene derivatives such as aniline and cresol has been studied by Gibbs, but these reactions have large temperature coefficients which make them less suitable for measuring the effective radiation than those possess- ing small or negligible temperature coefficients. In their tests at Manila they found that the average amount of oxalic acid decomposed per hour was 12.45 per cent, the minimum being 1.15 per cent and the maximum being 17.8 per cent. The average obtained by others at Baguio, Philippine Islands, was 14.9 per cent and at Honolulu, 13.9 per cent. According to Baudisch and Furst 49 the ammonium salt of alpha-nitrosonaphthylhydroxylamine turns red under exposure to blue, violet, and other radiations which pass through glass. If a piece of spongy paper is treated with this salt, then steamed and exposed to the radiation from a quartz mercury lamp, it will turn a reddish hue whether covered with glass or wholly exposed. If the paper, treated with a solution containing potassium nitrate or potassium iodide and starch, is exposed to the quartz mercury arc it will turn blue where it is bare but not where it is covered with glass. Bichromated gelatine is easily prepared and is quite useful in studying ultraviolet radiation. Details per- 186 ULTRAVIOLET RADIATION taining to this and many other preparations will be found in treatises on the subject. 50 Photographs of objects illuminated only by ultraviolet radiation differ in general in brightness distribution from that seen by the eye. Various so-called white objects, such as flowers and paints, do not necessarily appear of equal brightness in the photograph. There are many applications for this difference such as in the distinguish- ing between so-called black inks, the detection of erasures in checks and documents. Wood 51 made interesting photographs of the moon with ultraviolet filters before the camera. The craters of the moon, for example, ex- hibited new aspects. He described various screens. Michand 52 photographed twenty-four powdered alka- loids illuminated by ultraviolet radiation. He took one photograph of each under the usual conditions and one of each with a quartz lens silvered on both sides to ex- clude all radiation but a region between 300 and 330mpi. The photographs taken under ordinary conditions showed the alkaloids as white with the exception of that of ber- berine, which appeared black. The photographs made solely with ultraviolet radiation differed from the other group. Berberine still showed black, 12 were nearly black, 3 were gray, and 8 remained white. The phenomena of fluorescence and phosphorescence can be employed in the detection and measurement of ultraviolet radiation. In general, ultraviolet radiation excites photo-luminescence and it has been the author's experience that the near ultraviolet is usually most effec- tive. At least the middle region does not appear to be as effective in exciting phosphorescence in the sulphides as in producing photographic action when the effects of the near region are used for comparison. However, photo-luminescence is excited in many substances through- out the entire ultraviolet region represented in the radia- tion of the quartz mercury arc. DETECTION AND MEASUREMENT 187 Fluorescence and phosphorescence have been widely used for the purpose of detecting ultraviolet radiation and to some extent for actually measuring the intensity of radiation. During the recent war signaling by ultra- violet radiation was accomplished by directing the in- visible beam upon luminescent substances. This can be done by using a filter of dense cobalt glass which is trans- parent to near ultraviolet and opaque to visible radiation. Even more efficient filters can be obtained. One of the problems in such a case is to conserve the ultraviolet energy. It can be directed by a parabolic mirror and caught by another at a distance. In the first case the source of the radiation is at the focus of the mirror and in the second case the " luminescent " substance is at the focus. Various methods, sources, and filters were tried and as a consequence of combined experience such sig- naling was accomplished. A somewhat similar application was tried out in the convoy system and elsewhere. For example, a quartz mercury arc enclosed in a very dense cobalt blue glass was hung on one vessel. Observers on other vessels equipped with telescopes with fluorescent eye-pieces kept their vessels in correct positions by noting the position of the fluorescent image. It is known that infra-red radiation quenches phosphor- escence. Ives and Luckiesh 53 studied this phenomenon quite extensively. They also discovered that infra-red caused a momentary flashing-up of the phosphorescence of zinc sulphide several minutes after excitation. These phenomena can be utilized in signaling by projecting a beam of invisible infra-red upon glowing zinc sulphide. Hauer and Kowalski 5 * devised a monochromatic ultra- violet illuminator and a spectrophotometer designed to measure the comparatively feeble luminosities of phos- phorescent and fluorescent substances. They could study the effect of wave-length of the exciting radiation. They 188 ULTRAVIOLET RADIATION found, for example, that the fluorescence of lithium plati- nocyanide is a maximum when the wave-length of the exciting radiation is 390mpi. They found that the momen- tary and the enduring phosphorescence could be easily separated and, for example, the enduring phosphores- cence of phenanthrene is excited only by the radiations in the region of selective absorption of the substance. The rate of decay of phosphorescence depends upon the temperature, as Ives and Luckiesh 53 showed in their in- vestigations. Phosphorescence may consist of several spectral bands and if the rates of decay of these bands differ there is necessarily a change of color during decay. The author has observed different colors of phosphores- cences depending upon the wave-length of the exciting light. This was observed by focusing a large image of the ultraviolet spectrum of mercury upon phosphorescent and fluorescent substances. It appeared that the color- difference was generally due to the superposition of fluo- rescence (of slightly different color) upon the phosphor- escence. The observations of color-differences between the various images of the mercury lines upon the lumines- cent substance were generally made during excitation. Interesting changes in color during decay of phosphores- cence can be produced by mixing two luminescent sub- stances emitting, for example, red and blue phosphores- cence respectively. During excitation the combined color is purple but owing to different rates of decay the color may change during decay toward red or blue. Winther 55 has described a fluorometer, whose essen- tial principle has been used by various investigators. By means of this device the energy of a given wave-length can be measured in terms of the radiation of a standard lamp whose spectral energy-distribution is known. A pencil of radiation from the standard lamp is permitted to enter a quartz vessel containing a fluorescent liquid. This fluorescent beam is compared in brightness with DETECTION AND MEASUREMENT 189 the second or " unknown " beam which is admitted paral- lel and close to the first. Details of varying the inten- sity of the standard and of making the photometric comparison are obvious. Of course, the two radiations must have the same wave-lengths in order that their intensities may be proportional to their fluorescence. For the fluorescent materials he used solutions of, (1) rhodamine-B, 0.004 gram per liter (useful for wave- lengths shorter than 340m[i and for those between 460 and 600mfi) ; (2) sodium fluorescein, 0.01 gram per liter and 2.5 cc. of N-sodium hydroxide (useful between 254 and 520mpi); (3) quinine sulphate, 0.1 gram per liter and 4 cc. N-sulphuric acid (useful from 260mpi to the visible region). The phosphoroscope is a device for observing the phos- phorescence of a substance at any desired interval after excitation. It usually involves the rotation of the phos- phorescent substance and an adjustment of the observa- tion orifice so that it can be placed at any desired time- interval after excitation. For so-called fluroescent sub- stances the rotation is rapid because the period of decay of the luminescence is short. For phosphorescent sub- stances the rotation is slower and sometimes may be very slow. For example, Ives and Luckiesh 53 devised a phos- phoroscope by using an 8-day clock and placing a disk containing the material to be studied on the spindle of the hour-hand. They made single photographic exposures requiring as long as 240 hours to obtain the spectrogram desired. The phosphoroscope has been widely used by investi- gators of phosphorescent phenomena. Andrews 56 has described a phosphoroscope which includes a source of ultraviolet radiation and small motor for revolving a disk upon which the luminescent material is placed. A re- volving shutter eclipses the exciting radiation several thousand times per minute. By altering the speed of the 190 ULTRAVIOLET RADIATION motor the periods of exposure and of darkness may be changed. By placing a fluorescent screen in the plate-holder of a quartz spectrograph a great deal of qualitative data per- taining to transparency of media in the ultraviolet region can be obtained. An iron arc or other powerful source is focused on the slit by means of a quartz lens and the spectrum is viewed on the fluorescent screen. Uranium glass is very satisfactory for this screen. If stray light is eliminated and the room is quite dark so that the eyes may be adapted to the faint brightnesses, much can be done without resorting to photography. For the study of opaque materials, a hole may be provided in the side of the camera near the acute angle made with the plate- holder. Fluorescent liquids can be utilized or studied to advantage by projecting the exciting spectrum upon their surfaces. They may be viewed from above or from the side. There are a great many variations of interesting modes of attack. Kriiss 5r among others has described a spectrophoto- meter for the ultraviolet region in which a fluorescent screen is used in the eye-piece. The principle employed in this instrument is extremely valuable in qualitative studies and it has some possibilities in quantitative work. To separate radiations differing considerably in wave- length advantage may be taken of the chromatic aber- ration of a simple lens. In other words, radiation of shorter wave-length comes to a focus nearer to the lens than radiation of longer wave-length. This expedient has been used in infra-red work to excellent advantage. For separating ultraviolet from visible radiation a quartz lens may be used to bring the rays of a spark to a focus. This focus will not be at a point, but considering all wave- lengths, it will be along a line. If a small hole in a sheet of metal be placed at a certain point along this line, the radiation passing through the hole will consist chiefly of DETECTION AND MEASUREMENT 191 radiation near the wave-length of that radiation which happens to be focused upon the hole. The radiation com- ing directly from the spark along the optical axis consists of all radiations so that beyond the hole at some distance a small shield must be placed to absorb this radiation which includes undesired radiations. The principle has often been applied in research in radiation. If a more detailed description is desired reference may be made to an article by Andrews. 58 The ultraviolet may be separated from the visible radia- tion by forming a spectrum with a quartz spectrograph, screening off the visible, and recombining the ultraviolet spectrum by means of a quartz lens. Chemically pure substances in general exhibit fluores- cence and phosphorescence only faintly. Different kinds of glass exhibit characteristic fluorescence. Rods of sodium hydroxide exhibit a reddish fluorescence and a greenish color when rapidly moved away from the exciting source. There are many applications of phosphorescence not only in the detection of ultraviolet radiation but also in the detection of substances. Almost any substance fluoresces to some degree at least and a great many have been studied. Space does not permit a discussion of these investigations. Nichols and Merritt have conducted such investigations for years and much of their work has been collected in a single publication. 59 Among the substances which they investi- gated are rhodamin, fluorescein, eosin, chorophyll, ura- nium glass, fluorspar, esculin, resorufin, sidot blende, Balmain's paint (calcium sulphide), willemite, and vari- ous other aniline dyes and salts. The colors of the luminescence of a few substances are as follows: calcite, red; barium sulphide, orange; fluores- cein and eosin, yellow; cadmium compounds, yellow; uranium glass, greenish yellow; willemite, yellow-green; some salts of salicylic acid, blue; calcium sulphide and 192 ULTRAVIOLET RADIATION some other compounds of calcium, violet; calcium tung- state, light blue; zinc silicate, green. Other colors are emitted by other substances and they can be obtained by mixture. Such mixtures are interesting because of the different rates of decay of the luminescence of the com- ponents. Certain blue fluorspars exhibit luminescence after exposure to radiation. It is said that the phosphor- escence of these media excited by heat emits ultraviolet radiation which is photo-chemically active. An aqueous solution of quinine sulphate fluoresces a violet-blue color. It is excellent for demonstrating to a large audience the existence of ultraviolet radiation. Baskerville 60 has noted the application of ultraviolet radiation in testing minerals. Certain minerals are un- affected, some fluoresce and others phosphoresce. Kun- zite was discovered with the aid of ultraviolet radiation. Spodumene specimens were found to be generally unaf- fected. He claims that the fluorescence of diamonds is an indication of genuineness. A crushed mineral may be separated into fluorescent and non-fluorescent portions and this has been resorted to for testing willemite con- centrates and tailings. Willemite is fluorescent but the gangue is not. If the tailings contain no fluorescent par- ticles the concentration process is known to be efficient. The color of the fluorescence is generally diluted or altered by the body color of the substance. This can be easily seen in oils and dye solutions. A tablet of soda salicylate may appear white or slightly bluish in daylight owing to its blue fluorescence but the latter is quite over- whelmed by the luminosity of the reflected radiation. Under a high-tension spark, which emits relatively much less visible but a great deal of ultraviolet radiation, the blue fluorescence of the soda salicylate is more prominent. Ordinary glass and quartz are readily distinguished by interposing between the soda salicylate and the source. The former absorbs the radiations which/ excite the blue fluorescence but quartz does not. DETECTION AND MEASUREMENT 193 Wolff 61 found certain specimens of dehydrated potas- sium carbonate exhibited a reddish luminescence but the purified salt did not. He concluded that the luminescence was due to potassium sulphide, and that ultraviolet radia- tion is an aid in detecting this compound in commercial potassium carbonate. Tiede 62 found all pure preparations of magnesium sul- phide fluoresced faintly under exposure to radiation from the sun or an arc lamp. Apparently magnesium sulphide is sensitive particularly to the longer wave-lengths for it did not appear to respond to ultraviolet, radium, or Ront- gen rays. Cathode rays caused it to fluoresce blue or red. Stark and Meyer 63 have discovered that many sub- stances exhibit fluorescence in the ultraviolet region. This makes it necessary to recast some of the theories which are based only upon visual observations. Stokes was one of the earliest investigators of fluores- cence and phosphorescence and from his work he enun- ciated the law that the emitted radiation is always of greater wave-lengths than those of the exciting radiation which is absorbed by the fluorescent substance. Kauffmann 64 suggested that the change in color of fluorescence with change in the solvent follows the same order as the change of color itself. The fluorescent band of a solid substance lies farthest toward the ultraviolet, then follow the solutions in indifferent solvents, then those in dissociating solvents. According to Baly's theory the fluorescence continues a phase ahead of the absorption. Stark has advanced the opinion that all substances possessing selective absorption are fluorescent. This is not based upon complete experimental evidence but there are many data which seem to fit the theory. Wasicky and Wimmer 65 have used ultraviolet radiation for illuminating cocoa for the purpose of distinguishing by means of a microscope the shell and nib tissue. They used a carbon arc and the well-known filter consisting of 194 ULTRAVIOLET RADIATION " uviol " (Jena) glass cells one of which contained a con- centrated solution of copper sulphate and the other a solution (1 : 12000) of nitrosodimethylaniline. The tissue of the shell in this case appeared brownish ; the nib tissue appeared bluish-violet. Ultraviolet radiation may be used in this manner owing to fluorescent effects or in the case of photo-micrography because of the greater re- solving power of the microscope for radiation of short wave-length. Of course, in the latter case only the radia- tion of greater wave-length than about SSOmjj, would be effective with a glass optical system. The photo-electric cell has been used in the detection and measurement of ultraviolet radiation as has been noted in other chapters. In fact, it appears to be the most sensitive device for this purpose, provided a suitable galvanometer or electrometer is available. Many scien- tific applications have been made of photo-electric phe- nomena during recent years. Allen ti6 and Hughes 67 have presented extensive discussions of the phenomena. Herz, in 1887, discovered that ultraviolet radiation incident upon a spark-gap caused a decrease in the voltage necessary for the passage of a spark. It was soon found that the effect was due to the emission of electrons and the formation of ions. Hallwachs 68 was the first to in- vestigate the phenomena systematically. He found that a clean piece of zinc when illuminated by ultraviolet radia- tion, lost the electric charge which it possessed. It was also found that an insulated body acquired a positive charge when illuminated by ultraviolet radiation. This is due to the emission of negative electrons. Many metals exhibit the photo-electric effect when illuminated by ultraviolet radiation. The electric current produced increases with the intensity of the radiation. Apparently under certain conditions the current is pro- portional to the intensity of radiation. This proportion- ality may not always exist under the conditions of the DETECTION AND MEASUREMENT 195 experiment so that it is well to determine the relation between current and intensity of radiation in any given case. Elster and Geitel 69 have been pioneers in this field of research. They have shown that electro-positive bodies such as potassium and sodium are photo-electrically active under visible radiation. Zinc and aluminum exhibit the effect under solar radiation. Rubidium is photo-electri- cally active even under the radiation of a carbon filament lamp. Various metals, liquids, and vapors exhibit the effect so that there are many substances to choose from in utilizing the photo-electric effect in the detection and measurement of ultraviolet radiation. Photo-electric cells can be purchased or they can be made by following the methods described in the references. Kreusler 70 was one of the earliest to use the photo- electric phenomenon for the purpose of measuring ultra- violet radiation. He employed a piece of clean platinum in a vessel containing hydrogen at about 200 mm. pres- sure. The piece of platinum which was charged nega- tively, was near another metal which was the anode. The current flowing from the cathode to the anode when the former was illuminated by radiation was taken as a meas- ure of the intensity of radiation. Of course, he ascer- tained the relation between current and intensity of radiation. Elster and Geitel 71 have described very sensitive cells consisting of colloidal potassium. The alkali metals are quite widely used for photo-electric cells. Hughes 72 has described a sodium cell which appears quite suitable to the measurement of ultraviolet radiation. Among other results he found that the ionization of the air by ultra- violet radiation begins at about ISSmji. The relation of the wave-length of radiation to the photo-electric effect differs for various metals. Sodium is most sensitive to yellow light and potassium to blue 196 ULTRAVIOLET RADIATION light. Most metals exhibit the maximum effect under ultraviolet radiation. Among others Ladenburg 73 has studied this phase of photo-electricity. Using a mercury arc as the source and a fluorespar prism he found that copper, platinum and zinc exhibited the maximum photo- electric effect in the region of 215m^. For equal inten- sities of radiation the photo-electric effect increased with decreasing wave-length throughout the ultraviolet re- gions. That the middle ultraviolet is more active than the near ultraviolet radiation in producing the photo-electric effect with such metals as aluminum, magnesium, and zinc is shown by the fact that solar radiation is not as active as radiation from sparks and arcs. The active rays in solar radiation are almost completely absorbed by glass. How- ever, even visible rays produce photo-electrons to some extent from aluminum. The velocity of the electrons emitted by a metal plate in a vacuum is proportional to the frequency (or wave- length) of the incident radiation. This is Ladenburg's law and it apparently holds, at least approximately throughout the ultraviolet region. Lenard 74 discovered a volume ionization in the gas surrounding the metal plate. This is independent of the Hallwach's effect exhibited by the metal. Lenard showed that the ionization of the gas was associated with the ab- sorption of the ultraviolet radiation by the gas. One of the latest papers on the photo-electric photo- metry of ultraviolet radiation is by Elster and Geitel. 75 They discuss improvements in the cadmium ultraviolet photometer. Gibson 76 has recently described a photo-electric method of photometry which may serve as an example of the pro- cedure. This is a null method which has been used very successfully. He discusses the utility of the photo-elec- tric photometer as compared with that of the Hilger DETECTION AND MEASUREMENT 197 sector-photometer for the ultraviolet region; and the spectrophotometer for the blue and violet regions. The potassium-hydride cell which is on the market exhibits a maximum activity for radiation of 460m^ in wave-length when used with an incandescent lamp and a glass prism. Coblentz 77 has presented a summary of the character- istics and methods of use of the photo-electric cell and also a valuable bibliography of the subject. Among the more recent investigations of the color-sen- sitiveness of photo-electric cells is that of Seiler, 84 but the sensibility curves were not obtained for the ultraviolet region. Thirty cells were studied including all the alkali metals and hydrides of Na, K, Rb, and Cs. It was found that as the atomic weight of the alkali metal increases, the maximum sensitiveness decreases, the resonance peak becomes broader, and the wave-length of maximum sen- sitiveness shifts toward the red. The author suggests that these changes may be associated with the increase in atomic volume. For glass cells containing argon at low pressure the wave-lengths of maximum sensitiveness were found to be as follows: Li 405 m\i, Na 419m[A, K 440m^, Rb 473m^, Cs 539mpi, NaH 427m^, KH 456mp,, RbH 481m[A CsH 540mpi. It was found that the substitu- tion of quartz for the glass shifts the maximum toward longer wave-lengths. No fatigue effect could be detected in these cells. According to Halban and Geigel 85 the photo-electric cell is applicable to the measurement of absorption of radiation from 300 to GSOmjj, with the gas-filled tungsten lamp and as far into the ultraviolet as 253mpi with the mercury arc. Barnard 86 has described a microscope designed to be used with ultraviolet radiation. Objects that exhibit little or no structure by ordinary transmitted light are seen to be highly organized when examined by ultraviolet radiation and the structure seen is in part dependent on the wave- 198 ULTRAVIOLET RADIATION length used. Objects examined by this method must be dealt with in the living state or at least under conditions such that no change takes place in their constitution, hence the ordinary methods of mounting can not be em- ployed. The method is, in effect, its own staining process, differentiation of structure depending on the difference in absorption in the ultraviolet. The organisms or tissue are placed in any suitable fluid which is transparent to ultraviolet radiation and the photograph is taken. The slides used are of fused quartz with the smallest possible amount of gelatine upon them. Kogel 87 has successfully utilized fluorescence in the photography of palimpsests. Under the illumination by ultraviolet radiation (334mji) of a quartz mercury arc, the parchment fluoresces but the erased writing remains al- most dark, though the old inks used sometimes contained sulphur compounds. Chemical alteration of the parch- ment or greasing generally does not weaken the contrasts. This fluorescent photography often brings out detail not disclosed by the other methods and has much improved the exploration of old manuscripts. The index of refraction of air has been studied by many investigators but usually the work has been done with white light or with one monochromatic radiation. Re- cently Meggers and Peters 88 made observations at spec- trum intervals of about 4mji from 220 to 900mpi. Com- plete sets of observations were made on dry air at atmos- pheric pressure and at temperatures of 0, 15, and 30 C. They found that the data are quite closely repre- sented by certain dispersion formulae of the Cauchy form. They used these observations in the construction of a table giving the corrections which must be applied to wave-lengths measured in air whose density is not normal. They also present a table of corrections for converting wave-lengths or frequencies measured in air to their values in a vacuum. DETECTION AND MEASUREMENT 199 The dispersion of hydrogen has recently been investi- gated by Kirn 89 with the result that he has tabulated the refractive indices between 185.4 and 546.2m|A. Duclaux and Jeantet 90 have recently described a method of treating ordinary plates so as to increase greatly the sensitiveness to the radiation of the shorter wave-lengths. They had need of plates sensitive beyond 190mji, and tried the procedure advocated by Schumann, but found it tedious and uncertain. Schumann plates are distin- guished by the small proportion of gelatin, and it was thought that this condition could be secured by degela- tinizing to a great extent ordinary plates. Trials of various methods, such as immersion in warm water, acid solutions, digestive enzymes, were without success, but a simple and satisfactory procedure was devised. The plate is placed horizontally in a dish with dilute sulphuric acid (one volume of the strong acid to ten volumes of water), and kept for four hours at room temperature (about 77 F.), the temperature being a little higher than this at the beginning and a little lower at the end. They are then removed to a dish in which they are washed by a very slow current of water, as the remaining gelatin is tender. Thirty minutes will be a sufficient washing. They are then dried, which requires but little time on account of the small amount of gelatin present. Plates thus treated retain a thin layer of emul- sion poor in gelatin and uniformly spread on the glass. This deposit is extremely sensitive to ultraviolet radia- tion, but is also very fragile, and the authors recom- mend that before developing the surface should be coated with a thin film of collodion, the plate being immersed in the developing bath before collodion is quite dry. Although most commercial plates are adapted fairly well for this procedure, it is likely that trial with many forms will show some more suitable than others. For rays of much greater wave-length than above noted, these plates 200 ULTRAVIOLET RADIATION are ten times more sensitive than the best plates prepared according to Schumann's method, and at least 200 times as sensitive as the plate in its commercial form. Another method for obtaining plates of high sensitive- ness to radiation of short wave-lengths is by covering the emulsion with a layer of fluorescent substance. Such a substance absorbs, so to speak, the short waves and emits in turn waves of greater length, to which the gelatin is transparent, and thus permits an action on the silver com- pound, hence the impression is made as if the gelatin was not present. For this method, substances giving blue or violet fluorescence should be chosen, and they should be dissolved in a liquid that will not swell the gelatin, and is not absorbed by it, since the efficiency of the process de- pends on the fact that the fluorescent rays act before the light enters the gelatin film. Water is, therefore, not applicable. Duclaux and Jeantet obtained good results with a solution of esculin in glycerol, but found most satisfactory results with lubricating oil. Many of the commercial forms of these have a distinct fluorescence due to hydrocarbons. It is sufficient to smear a few drops of such an oil over the emulsion by means of a wad of cot- ton. After exposure this film should be removed by means of ether or alcohol. A very thin fluorescent layer may be obtained by immersing the plate for a few minutes in a solution of the fluorescent oil in light petroleum or alcohol and allowing the solvent to evaporate. These procedures are simple and effective. They enable the operator to secure photographs of rays ranging from the extreme red to the limit of the ultraviolet. One slight defect is noted, a very small enlargement of the rays by irradiation, but this does not go beyond the twentieth of a millimetre. The processes have been tried with many commercial plates, and the sensibility is found to be greater than with the sulphuric acid method. It is possible, indeed, to carry out an instantaneous spectrography. Detailed results DETECTION AND MEASUREMENT 201 with certain metallic spectra are given in the original paper. In this chapter only glimpses of the chief facts of very extensive fields such as photography and photo-electricity have been presented. Some applications are touched upon in various chapters but for more complete treatises especially upon photography and photo-electricity the reader may consult the references given. It may appear that the photo-electric cell has not been adequately dis- cussed; however, it appears the better plan to wait until it becomes better developed and the procedure more standardized for every-day applications. References 1. Bui. Bur. Stds. 9, 1912, 7. 2. Outlines of Applied Optics, 1912. 3. Ann. d. Phys. 37, 1899, 529. 4. Ann. d. Phys. 35, 1910, 928. 5. Ann. d. Phys. 24, 1907, 25. 6. Phys. Zeit. 4, 1903, 614 and 861 ; 5, 1904, 34. 7. Phil. Trans. 166, 1876, 325. 8. Phys. Rev. 4, 1897, 2 97- 9. Bui. Bur. Stds. 4, 1908, 404. 10. Proc. Roy. Soc. 42, 1887, l8 9J 44 l888 9 6 ; 47> l8 9 4 8 - 11. Soc. Franc, d. Phys. 1886, 30 and 77. 12. Soc. Franc, d. Phys. 1908, 148. 13. Ann. d. Phys. 29, 1909, 1003. 14. Bui. Bur. Stds. 2, 1906, 476. 15. Proc. Roy. Soc. 23, 1910. 16. Arch. Sci. Phys. Nat. 18, 1887, 347. 17. Color and Its Applications, 1921, 200. 18. Ann. d. Phys. 13, 1904, 890. 19. Phys. Zeit. 4, 1903, 614 and 861 ; 5, 1904, 215. 20. Ann. d. Phys. 8, 1901, I. 21. Smithsonian Contrib. No. 1413; Ber. Akad. Wien. 102, Ila, 625. 22. Astrophys. Jour. 23, 1906, 181. 202 ULTRAVIOLET RADIATION 23. Proc. Manchester Phil. Soc. 49, 1904, I. 24. Proc. Roy. Soc. 1919, 258. 25. Proc. Roy. Soc. Edinburgh, 32, 1912, 40. 26. Chem. Abs. 1917, 3133. 27. Smithsonian Physical Tables, 1920. 28. Wied. Ann. 45, 1892. 29. Ann. d. Phys. 6, 1901; 8, 1902. 30. Inaug. Dis. Berlin, 1909. 31. Ann. d. Phys. 12, 1903, 1155. 32. Trans. Farady Soc. 7, 1912, 267. 33. Phys. Zeit. 14, 1913, 336. 34. The Photographic Process, 1907. 35. Photochemistry, 1914. 36. Phys. Rev. 8, 1916, 674. 37. Bur. Stds. Tech. Pap. 148. 38. Comp. Rend. 156, 1913, 389. 39. Jour. d. Phys. 3, 1913, 305. 40. Phys. Rev. 16, 1903, 129. 41. Trans. I. E. S. 10, 1915, 149; Elec. World, 64, 1914, 129 and 954. 42. Ann. d. Phys. 41, 1913, 1012. 43. Astrophys. Jour. 43, 1916, 302. 44. Ann. d. Phys. 5, 1901, 349. 45. J. Opt. Soc. July, 1917. 46. Phys. Rev. 17, 1921, 246. 47. Chem. Ztg. 34, 267. 48. J. Phys. Chem. 16, 1912, 709. 49. Ber. 45, 1912, 3426. 50. Cassell's Cycl. of Photography. 51. Smithsonian Inst. 1911, 155. 53. Astrophys. Jour. 34, 1911, 173; 36, 1912, 330. 52. Science, 1912, 415. 54. Phys. Zeit. 15, 1914, 322. 55. Z. Elecktrochem. 19, 1913, 389. 56. Geij. Elec. Rev. 1917, 259. 57. Zeit. f. Inst. 23, 1903, 197 and 229. 58. Gen. Elec. Rev. 1917, 817. 59. Carnegie Inst. Pub. No. 152. 60. Electrochem. Met. Ind. 1906, 435. DETECTION AND MEASUREMENT 203 61. Chem. Zeit. 36, 1912, 197. 62. Ber. 49, 1745. 63. Phys. Zeit. 8, 1907, 250. 64. Ber. 41, 1908, 4396. 65. N. Nahr. Genussm. 30, 1915, 25. 66. Photo-Electricity, 1913. 67. Photo-Electricity, 1914. 68. Wied. Ann. 33, 1888, 301. 69. Ann. d. Phys. 38, 1889, 4 an d 497- 70. Ann. d. Phys. 6, 1901, 412. 71. Phys. Zeit. 12, 1911, 758. 72. Phil. Mag. 25, 1913, 679. 73. Phys. Zeit. 8, 1907, 590. 74. Ann. d. Phys. i, 1900, 486; 3, 1900, 298. 75. Phys. Zeit. 1915, 405. 76. Bur. Stds. Sci. Pap. No. 349. 77. Bur. Stds. Sci. Pap. No. 319. 78. Astrophys. Jour. 52, 1920, 47. 79. Astrophys. Jour. 53, 1921, 150. 80. Proc. Roy. Soc. 99, 1921, 78. 81. Collected Works, vol. i, p. 213. 82. Phil. Trans. A, 217, 1917, 241. 83. Comp. Rend. 170, 1920, 226. 84. Astrophys. Jour. 52, 1920, 129. 85. Z. Phys. Chem. 96, 1920, 214. 86. Nature, 106, 1920, 378. 87. Preuss. Akad. Wiss. Berlin, Ber. 37, 1914, 974. 88. Bull. Bur. Stds. 14, 1918, 697. 89. Ann. d. Phys. 64, 1921, 566. 90. Jour. d. Phys. ii, 1921, 156. CHAPTER XI EFFECTS UPON LIVING MATTER That radiation affects living cells is evident by sun- burn, snow-blindness, sterilization, and in many other ways. Man has no sensory organs for detecting radi- ation beyond the limits of the visible spectrum. He protects himself from the glare or the heat of the sun which saves him from the slight amounts of harmful ultraviolet radiation. Fortunately the short- wave limit of the spectrum of solar radiation is about the same as the short-wave limit of transparency of the cornea. As a matter of fact the eye evolved through adaptation to solar radiation and, therefore, it is not strange that these two limits practically coincide. Man has devised artificial sources of ultraviolet radiation of such wave-lengths as to be harmful. When these invisible rays accompany the light rays much damage may be done. But with the development of such sources, knowledge of the effects of radiation increased so that man is able to protect himself from the harmful rays and also to utilize them to his advantage. It has long been known that intense radiation, especially of the shorter wave-lengths, caused painful irritation of the anterior tissue of the eye. " Snow-blindness " is a common result of intense solar radiation and it is now known that snow reflects the ultraviolet rays in solar radiation very efficiently. The result is a painful irritation which usually becomes apparent several hours after ex- posure. The eyes first appear to contain foreign matter, that is, to feel "sandy." This disorder has been termed " photophthalmia." Various associated effects arise from 204 EFFECTS UPON LIVING MATTER 205 gazing at a partial eclipse of the sun and this disorder has been termed " eclipse blindness." Since the advent of artificial sources rich in ultraviolet radiation, the first one being the carbon arc, many cases have arisen of what was termed " ophthalmia electrica." It has long been known that the effective radiation has been that of the shorter wave-lengths in solar radiation and also those in the middle and extreme regions of the ultraviolet from artificial sources. Verhoeff and Bell 1 in their extensive investigations, which will be referred to occasionally, found that it is the radiation shorter than 305mjx in wave-length which is able to injure cells by chemical action. They found that at least 2 x 10 6 erg-seconds per sq. cm. of such energy is necessary to produce a well marked photo- phthalmia. They found that for any source yielding rays capable of producing pathological effects on the cornea, the time of exposure required to produce the symptoms of photophthalmia is inversely proportional to the in- tensity of the radiation of the effective rays. Of course, allowance must be made for " physiological repair " when the intensity of radiation is so feeble as to require very long periods of exposure. They verified the inverse- square law over a large range of intensities of radiation, that is, that the time required for the development of characteristic symptoms varied according to the inverse square of the distance from the source of the harmful radiation. They also found that the energy effects are additive for at least the first 24 hours of intermittent exposure. According to these investigators there is a practical limit to the abiotic action of ultraviolet radiation, for the action of radiation of longer wave-length than SOSm^ is so slight as to be overcome under ordinary circumstances by the physiological activities of the cells. Hallauer 2 found that the lens of the adult human eye 206 ULTRAVIOLET RADIATION is opaque to radiation shorter than 376my, in wave-length and often to that shorter than 400m^. He found a slight transparency between 315 and 330mpi in some lenses from the eyes of children. It may be safely stated that the retina of the average adult eye does not receive radiation of shorter wave-length than SSOm^ and not much energy of shorter wave-length than 400m[i. Parsons 8 studied the spectral transmission of the parts of a rabbit's eye in normal saline solution. He found that the absorption by the lens began at 400mji and became complete at 350mpi. This result has been obtained by various other investigators. There is a great deal of evidence that the transparency of the cornea extends to 295mpi; that is, the cornea trans- mits radiation of greater wave-length than 295mjj,. Par- sons found a layer, 3/16 inches thick, of the vitreous humor of a rabbit's eye, to be transparent to 280mji, and the transparency to extend to 270mji. He also found that measurements upon eye-media several hours after death yielded results identical with fresh specimens. Schanz and Stockhausen 4 have published the trans- mission spectra of eye-media. They found that the cornea was transparent to 300m[x and later decided that it was transparent only as far as 320m|i for practical considera- tions because the transparency rapidly diminished from 320 to 300mji,. Martin 5 obtained results which agreed with those of Parsons, that is, that the cornea was trans- parent to 295 mjju In general, the aqueous and vitreous humors transmit radiation of shorter wave-length than that transmitted by the lens. Birsch-Hirschfeld 6 found the limit of trans- parency of a thickness of 1 cm. of vitreous humor to be at SOOmjj, and that it was the same for all animals. This result has been confirmed by others. There appears to be a considerable variation in the transparency limit of lenses from the eyes of different EFFECTS UPON LIVING MATTER 207 animals as well as of lenses of the same species. As already seen, the latter statement applies to human lenses. Widmark 7 found that the short-wave limit of visibility of radiation changed with age. Children between the ages of 1 1 and 20 years obtained a visual sensation from radia- tion of wave-length as short as 3S6m\i. The limit short- ened with increasing age so that it was at 402mji for per- sons between the ages of 62 and 74 years. Of course, the limit is not necessarily established by the transparency of the lens. Apparently, fluorescence of the lens is caused by radiation between 350 and 400mji in wave-length and the maximum effect is due to radiation of wave-length 385mpi. It is quite evident that the lens of the human eye ab- sorbs the radiation between 295 and 350mfi in wave-length which is incident upon the cornea. An evidence of its absorption is the fluorescence which it exhibits. No radiation of shorter wave-length than SSOmji and perhaps 380mpi, can reach the retina of an adult eye. In fact, as the eye ages, the cornea becomes yellowish and its ab- sorption extends sometimes only as far as 420mpi. Chardonnet 8 employed a silvered quartz plate of such thickness of silver film as to be opaque except for the region between 301 and 343m^ in attacking the question of transparency of eye-media. Normal eyes could not see an electric arc through this glass but eyes from which lenses had been removed through operation for cataract, could detect movement of the arc. It is certain that the radiations most effective in causing sunburn, irritation, and, in fact, the destruction of animal tissue, are chiefly confined to the middle ultraviolet region. Schunck 9 exposed the forearm to ultraviolet radiation and concluded that the greatest effect was in the region of 235 to 250mpi although the region was indefinite. Henri and Moycho 10 exposed the ear of a rabbit to the ultraviolet spectrum from 230 to 330m|x. The most 208 ULTRAVIOLET RADIATION active radiation was that at 280mpi and the energy neces- sary to produce irritation was found to be 0.057 x 10 7 ergs per sq. cm. No effects were detected for 330 m^i and for the region of shorter wave-length than 250 m|x. The work of Verhoeff and Bell x has shown that abiotic action for living tissues is confined to wave-lengths shorter than 305mjA. Their work was exhaustive and it included a discussion of the investigations by others. Their pub- lished article is accompanied by an extensive bibliography and digest of the pertinent literature by Walker. 11 Burge 12 found that the radiation from a quartz mercury arc sufficiently intense to coagulate egg albumen, egg globulin, vitellin, serum albumen, and serum globulin in one hour of exposure, did not coagulate the protein in the normal lens or of the humors in 100 hours of exposure. The radiation which coagulated egg-white was between 265 and 320mpi, the most effective being near 265m^. The lens protein can be modified by solutions of calcium chloride, magnesium chloride, sodium silicate or dextrose too weak of themselves to affect the transparency of the lens, so that ultraviolet radiation can precipitate the modi- fied protein and thus produce opacity of the lens. The effective region was between 265 and 302mpi, for the source of radiation employed. He claims that senile cata- ractous human lenses show that calcium, magnesium, and in lenses from India, silicates are greatly increased in this type of cataract. The assumption is made that the ac- cumulation of these substances modifies the lens protein in such a way that the ultraviolet radiation precipitates the protein thus producing cataract. According to Burge 13 ultraviolet radiation kills living cells and tissues by changing the protoplasm of the cells in such a way that certain salts can combine with the protoplasm to form an insoluble compound or coagulum. He found the effective radiation to be between 254 and 302mji in wave-length. He produced cataract in the eyes EFFECTS UPON LIVING MATTER 209 of fish living in dilute solutions of those salts, found to be greatly increased in human cataractous lenses, by ex- posing the eyes to the radiation from a quartz mercury arc. Abnormal quantities of the salts of calcium and sodium silicate in the cells of the eyelids and of the cornea increase the effectiveness of ultraviolet radiation in pro- ducing anterior eye trouble. Abnormal quantities of these upon the skin also increase the effectiveness of the ultraviolet rays in solar radiation in producing sunburn. Many germs appear to thrive better in the dark than when exposed to solar radiation. Perhaps ordinary visible radiation kills some kinds of germs but in general it is ultraviolet radiation which is effective. The modern artificial sources, such as the quartz mercury arc, the carbon and flame arc, the magnetite arc, and such pro- cesses as arc welding, yield ultraviolet radiation of powerful germicidal action. Henri and his wife 14 found that the absorption of egg- albumen for radiation of various wave-lengths cor- responded closely to the time-value of bactericidal action for the same radiations. The absorption and abiotic action are taken to be indicative of each other. If this is true, the abiotic action is powerful near 200m|A and rapidly diminishes for radiation between wave-lengths 210 and 230mpi. From 250 to SlOmpi the effect is rela- tively small and decreases slowly. It apparently ends at SlOmjLi. They found the abiotic action at 215mpi to be about 25 times greater than at 250mpi and several hun- dred times greater at SOOmpi. In experiments which aim to determine the effect of ultraviolet radiation upon animal tissue, the so-called " heat effect " must be eliminated. This is especially true if the radiation is focused upon the tissue by means of lenses. Obviously, the radiation is destructive if it is of sufficient intensity to burn in the ordinary sense. A water-cell usually serves satisfactorily enough for eliminating a great deal of the infra-red. 210 ULTRAVIOLET RADIATION Browning and Russ 15 have described experiments with radiation from 210 to 700mpi recorded by means of a quartz spectrograph. A glass plate coated with nutrient agar and an emulsion of staphylococcus pyXogenes aureus was exposed in the spectrograph for several minutes. It was then incubated and it was found that germicidal action had taken place between 238 and 294m|4,. The maximum action was between 254 and 280mji. An ex- posure of 3^ hours showed the limits of germicidal action to be 215m|A and 296mpi. Radiations between 296 and 380mji exhibited no effect on germs but they were found to penetrate the human skin for an appreciable depth. The radiation between 210 and 296mji was absorbed by a thickness of 0.1 mm. of human skin. The more or less indefinite long-wave limit of germi- cidal action is in the vicinity of the short-wave limit of the solar radiation which reaches the earth. It is likely that a slight change in atmospheric conditions or a large variation in intensity of radiation might have a great influence upon the germicidal action of solar radiation. Bo vie 16 studied the germicidal power between 250 and SOOmji. He found that radiation of wave-lengths shorter than 292.5m\i killed bacteria and spores of vari- ous fungi in 10 minutes but radiation 2.5m^ longer in wave-length did not kill in two hours. In Chapter II it has been seen that the end of the solar spectrum varies in the vicinity of these wave-lengths. Bovie found that ultraviolet radiation can penetrate blood-filled tissue to a depth of only a fraction of a milli- meter, but if the skin is rendered anemic by eliminating the blood by pressure, ultraviolet radiation kills bacteria through 4.25 mm. of tissue. He discussed at consider- able length the interesting phase of fluorescence in re- lation to germs, living tissues, and various chemical reactions. Plate X. Radiant energy is finding many applications in therapeutics. EFFECTS UPON LIVING MATTER 211 Burge 17 studied the bactericidal action of the radia- tion from the quartz mercury arc upon seven different kinds of non-fluorescent bacteria and upon eight dif- ferent types of fluorescent bacteria. He found that an exposure of 200 seconds killed all the non-fluorescent bacteria but at the end of that time none of the cultures of fluorescent bacteria were completely killed. Cernovodeanu and Henri 18 investigated the bactericidal action of the ultraviolet radiation emitted by mercury arcs. They employed emulsions containing from 10000 to 100000 bacteria per cubic centimeter and found that the action decreased more rapidly than the inverse square of the distance from the source. They found that air, oxygen, or hydrogen-peroxide was not essential to the destruction of bacteria by ultraviolet radiation. The most destructive radiation was found to be in the vicinity of 280mj,i. The resistance of different species of bacteria varied, bacillus coli being killed in 15 to 20 seconds and bacillus subtilis in 30 to 60 seconds. They found that protoplasm absorbs ultraviolet radiation shorter than 290mj.i in wave-length. The same investigators using as a standard the action of radiation which passed through glass (wave-lengths greater than 302mpi) found the bactericidal power of radiation passing through a viscous screen (wave-lengths greater than 253mpi) to be 300 times greater than that of the radiation passing through the glass but only 1.6 to 5 times greater in chemical action. The bactericidal power of radiation passing through a quartz screen was about 1000 times greater than the standard but the chemical action was only 4 to 6 times greater. A mer- cury arc was the source of radiation and the chemical reactions were the decomposition of hydrogen peroxide, the decomposition of potassium iodide in the presence of hydrochloric acid and of starch, the reaction between mercuric chloride and ammonium oxalate, the blackening 212 ULTRAVIOLET RADIATION of silver citrate paper, and the oxidation of leuco-deriva- tives of fluorescein. Recklinghausen 19 has discussed the sterilizing action of ultraviolet radiation from the quartz mercury arc. For the sterilization of water it should be free from sus- pended particles which cast shadows and thereby shield the bacteria. The water should be agitated. Colloidal material and coloring matter in small quantities does not seriously reduce the effectiveness. Clear ice does not interfere. He states that 100 kilowatt-hours of electric energy will sterilize one million gallons of water. There is an obvious advantage over the introduction of chem- icals because of the unsatisfactory flavor caused by the latter. He 20 has discussed the economics of sterilization of water by ultraviolet radiation from the quartz mercury lamp. The quartz tube is made of the pistol type which is inside a quartz cylinder projecting into the water. He describes large lamps operating on a 500-volt circuit at 3 amperes and claims that the production of ultraviolet radiation is 50 times greater than in the case of the 110- volt, 3.5-ampere lamp with a consumption of energy only 4 times as great. One lamp will sterilize about 1000 tons of water per 24 hours. The same investigator 21 has stated that water after leaving a mechanical filter can be sterilized at a total cost of 60 cents per million gallons. A comparison with ozone showed that ultraviolet radiation and ozone steri- lized water at about the same cost for power but that the initial and operating cost of ozone equipment was higher than for the quartz mercury-arc apparatus. A patent 22 has been issued to von Recklinghausen which describes an apparatus for passing the liquid to be sterilized in thin films before a quartz mercury arc at various points. It involves the disposition of baffle plates. Mineral deposits are formed on the quartz pro- EFFECTS UPON LIVING MATTER 213 tecting tube which absorb the ultraviolet radiation. These are eliminated by circulating water around the tube. Opaque or colored liquids are sterilized by passing through a shallow chamber in thin film. Barrels have been sterilized by inserting a quartz mercury arc through the bung-hole. / ( The spectra of blooof-coloTmg matters such as oxy- haemoglobin, haemoglobin, carbonyl-haemoglobin, me- thaemoglobin, acid haematin, alkaline haematin, haem- ochromogen, acid haematoporphyrin, and alkaline haem- atoporphyrin all exhibit absorption in the near ultra- violet. It has also been shown that red blood corpuscles are destroyed by radiation less than SlOmpi in wave-length. Deblet and Beauvy 23 studied the effect of ultraviolet radiation in the hemolytic power and colloidal state of blood serum. After about an hour's exposure it was found that the hemolytic power of the blood serum was reduced about one-half. After 23 hours no changes were visible with the aid of the ultramicroscope. Berthelot claims to have reproduced the digestive processes by the use of radiation from a quartz mercury arc and without the aid of ferments which are important in the natural process. Henri and Moycho 24 determined the action of ultra- violet radiation of various wave-lengths upon tissue and also the radiation responsible for sunstroke. According to Grimm and Weldert, 25 who employed a quartz mercury arc of 1200 candle power, clear water containing less than 100 bacteria per cubic centimeter was sterilized at a rate of 0.55 cubic meters per hour but in cases of water containing many more bacteria the rate of passage of water through the apparatus did not exceed 0.45 cubic meters per hour. They considered the cost of sterilization to be excessive compared with other existing processes. Small sterilizers are available which employ the quartz 214 ULTRAVIOLET RADIATION mercury arc. Swimming pools are effectively purified by units of moderate size. Air bubbles by causing tur- bidity are the cause of incomplete sterilization. By exercising care the path of the water can be such as to minimize the production of bubbles. Henri 2e found that the bactericidal power of ultra- violet radiation is proportional to the coefficient of absorp- tion of protoplasm, thus indicating that the action of ultra- violet radiation on micro-organisms obeys the common law of photochemical absorption. This also suggests that the destruction is brought about by direct action upon the cells instead of indirectly through the formation of hydro- gen peroxide and other chemicals. The radiations of greatest bactericidal power penetrate only a few thou- sandths of a millimeter. The penetration of solar radiation into the human flesh is reduced by acclimatization by the formation of a pro- tective layer of pigment which is more or less opaque. This acclimatization may be either permanent or tem- porary. It has become permanent through ages of adaptation as is evidenced by the black or dark-colored races of the tropics. The tanning of white skin is a temporary effect which may become more or less per- manent. The skin of the white race contains a slightly protective pigmentation which varies to some extent and this variation is responsible for the differences in sensi- tivity to solar radiation. For example, blondes who possess only very slight pigmentation are more easily sunburned and are more susceptible to sunstroke. Courmot 27 found that liquids containing suspended colloid particles absorb ultraviolet radiation strongly. He claims that the sterilization does not involve ozone or hydrogen peroxide. He investigated wine, beer, peptone solution and other liquids. Other investigators have found that malt liquors cannot be sterilized effectively by ultraviolet radiation owing to their opacity for the effective rays. EFFECTS UPON LIVING MATTER 215 Burge 28 has discussed the mode of action of ultra- violet radiation in effecting sterilization. Various applications have been made of ultraviolet radiation in the sterilization of milk. One 2 * involves the flow of milk, beer, water, and other liquids from per- forated pipes over inclined quartz plates and corrugated metal plates. These plates are suitably arranged in re- spect to the mercury arc. Another process 30 involves freezing the milk into blocks or sheets and exposing them to the radiation. As the frozen milk thaws the thin film of liquid is sufficiently transparent to permit the radiation to penetrate. Another process 31 combines the use of heat and ultraviolet radiation. It is claimed that by heating the milk to 60 C the bacteria are so enfeebled that they succumb to a very slight exposure to ultraviolet. It has been claimed by Ayers and Johnson that ultra- violet radiation does not affect vegetable cells and that this treatment cannot be substituted for pasteurization because there would be no certainty regarding the de- struction of pathogenic germs. According to them, ultra- violet radiation causes a disagreeable taste and odor in milk. Butter and other fats have been sterilized by ultra- violet radiation. In one method 32 the fat is spread in a thin layer on an endless belt which passes under sources of ultraviolet radiation. Bovie 33 investigated the action of the radiation from a quartz mercury arc on living cells through its effects on the constituents of protoplasm. Ox-serum was coagulated in quartz tubes but not in glass ones. Fresh egg albumen was slightly coagulated in two hours at a distance of 10 cm. from the arc. He also studied the influence of tem- perature. Doree and Dyer 34 exposed clean bleached cotton cloth to a mercury vapor (glass tube) lamp for a week. They concluded that the ultraviolet radiation converted cellulose into oxycellulose and it lost its tensile strength. 216 ULTRAVIOLET RADIATION Chauchard and Mazoue 35 found that the two enzymes, amylose and invertase, suffered the loss of their activity when exposed at a distance of 12 cm. from a quartz mer- cury arc. Agulhon 36 has concluded that enzymes may be divided into three classes according to their sensitiveness to radia- tion. Sucrase, tyrosinase, and laccase are destroyed by visible radiation in the presence of oxygen but in a vacuum only ultraviolet radiation is destructive. Emulsion and catalase are destroyed by radiations of all wave-lengths in vacuo or in air. Oxygen hastens the destruction. He gives rennet as an example of the third type. Its activity is destroyed in a vacuum by ultraviolet radiation and is diminished by visible radiation. Chauchard 37 investigated the effect of wave-length of ultraviolet radiation on the destruction of amylase and lipase using sparks of zinc, cadmium, and magnesium. Radiations between 220 and 250mji were more destructive than those of longer wave-length in the case of amylase. He 38 later found that the amylase of pancreatic juice was appreciably impaired only by radiation of about 280m|x in wave-length. Gamgee 39 has presented an account of the investigations of the absorption of violet and ultraviolet radiation by haemoglobin and its derivatives. Various researches on the absorption spectra have been made. Normal blood of rabbits, sheep, and pigs possesses the same absorption spectra. The maxima of the bands of methaemoglobin are slightly displaced in comparison with the oxyhaema- globin bands. According to Schumm 40 the absorption spectrum of oxyhaemaglobin exhibits three bands, one of which is in the ultraviolet. He states that some variations appear in the spectra of blood of the same species and that the oxy- haemaglobin of different animals cannot be distinguished in this manner. Mashimo 41 found the maximum of the EFFECTS UPON LIVING MATTER 217 ultraviolet band of oxyhaemaglobin to lie at 35 Om^ and was unable to observe any band at the extreme region of ultraviolet. It has been found that diphtheria toxin is rendered atoxic by ultraviolet radiation. According to Lowenstein, the quartz mercury lamp is 30 times more effective than the iron arc. A 40-hour exposure to it completely de- stroyed 1000 fatal doses of the toxin. In his experiments radium had no destructive influence. Ultraviolet radiation has been used very extensively in therapeutics, but the results which have been described are more or less confusing and indefinite owing to the absence of data pertaining to the spectral character. Woodruff 42 has accumulated many abstracts pertaining to actino-therapy. Ultraviolet radiation is said to relieve pain from super- ficial new growths and to lessen the odor of putrefaction. Primitive beings are known to expose a wound to the direct radiation from the sun with apparent hastening of healing. Solar radiation is said to cure persistent bed- sore and many skin diseases. However, there are many claims which are contradictory. Certainly there are many cases where ultraviolet and visible radiations are harmful, and this has led to some applications of red light not for its effect but rather for its absence of effect. Sun-baths which were quite the thing some years ago now have many opposed to their use. Owing to the contradictory nature of many opinions, to the lack of physical data pertaining to the radiation employed and to the lack of control of experiments in many cases, the therapeutic aspects can not be discussed without many digressions and qualifica- tions. To do this appears out of place here so it is recommended to those particularly interested that they consult the book of Woodruff 42 and the one by Cleaves. 43 Ultraviolet radiation has been detected at a depth of 1000 meters in the ocean and blue and violet rays at a 218 ULTRAVIOLET RADIATION depth of 500 meters. Apparently no radiation penetrates to depths of 1700 meters according to Hjort 4 * whose in- vestigations were made in the north Atlantic. Radiation necessary for the growth of vegetation does not pene- trate in sufficient quantity beyond the depth of 400 to 600 meters. There appears to be little doubt that powerful solar radiation sterilizes water, for more bacteria and vegetable plankton are found in winter than in summer in various waters. They are also more plentiful in foggy climates than in regions of more clear weather. Insects suffer from ultraviolet in proportion to their lack of protection. The white ant cannot withstand ex- posure to solar radiation and therefore is obliged to con- struct subways in search of food. Many kinds of insects are easily killed by the radiation from the quartz mer- cury arc. Solar radiation is fatal to some kinds of mos- quitoes, hence they avoid it and are most active at dusk and during the darkness of night. The grubs of wasps and bees are killed by solar radiation, hence they are protected in the cells which are light-proof. Woodruff 42 believes that it is strange that the human being does not have a nerve apparatus to receive impres- sions from ultraviolet radiations intense enough to be harmful or useful and still too feeble to make their pres- ence known by conversion into heat. He suggests that our remote ancestors evolved in dark cloudy regions where ultraviolet radiation existed only in small quantities. Hence, no nerve-sense evolved as there was no need of it, but evolution later turned toward the development of protective pigment. Then it was too late to evolve by variation a new nerve sense. It is reasonable to suppose that plant life as well as animal life has become particularly adapted to the radi- ation to which it has become accustomed throughout cen- turies of evolution. The primitive savage who lives un- clothed is not subject to sunburn as is the civilized being EFFECTS UPON LIVING MATTER 219 who has lowered the resistance of his skin by the use of clothes. The cornea of the eye transmits practically all the short-wave energy of solar radiation. In fact, animal cells are not affected by solar radiation except under con- ditions which are uncommon. Plants also flourish in their customary environment whether in the desert, on the mountain top, or in the cool shadows. Beyond the out- skirts of their ideal environment they show the scars of battle against hostile conditions. However, in general, it may be stated that solar radiation does not contain rays which kill plant life when the latter is in its customary environment. Many investigations have been made upon the effect of radiation on plants but most of the results are indefinite owing to the absence of specific knowledge pertaining to the spectral character. Furthermore, much systematic work must be done before it can be stated with certainty that ultraviolet radiation of certain wave-lengths produces certain results. Perhaps some rays act only as catalysers ; that is, they may have no direct influence but their presence is necessary. Of course, it is known that the development of chlorophyll is a photo-chemical reaction and its absorption spectrum is well known. Shanz 45 has attempted to prove that all organic sub- stances appear to be altered by the radiation which they absorb. He claims to have proved that organic sub- stances are broken up by radiation into their elements and radicals. In the case of colorless substances the effective radiation is chiefly the ultraviolet and this is supposed to be true for those substances which are unaffected by visible radiation. According to him, the shorter the wave- length of radiation the greater the power of breaking down the structure of the molecules. Radiation produces some more striking effects upon plants than upon animals. The effect produced upon plants by radiation is most obvious in the process of assim- 220 ULTRAVIOLET RADIATION ilation. The chlorophyll grain and the colorless struma, the chromoplast, are the principals. The former is a fluorescent dyestuff and the latter is albumen which Schanz considers to be sensitive to radiation of short wave- length. The cell sap penetrates the chlorophyll grain and carries to the latter the materials it requires in the process of assimilation. Among these materials there are sup- posed to be some which influence the photo-chemical re- action in the manner of catalysers. Schanz considers that the chief function of the colors of flowers is to select the radiation required in any particular case and in accordance with this selection peculiar sub- stances are carried over with the seeds to the new indi- vidual. The concept of colors as sensitizers implies that they are highly important with regard to the plant's own needs. He claims that assimilation is caused chiefly by radiation of the longer wave-lengths ; that is, by those rays to which albumen is not sensitive in itself and to which it must first be sensitized, accordingly, by means of chloro- phyll. The radiation of shorter wave-lengths appears to take less part in this process, notwithstanding the fact that they are otherwise chemically more effective. This led Schanz to inquire why ultraviolet radiation plays such a small part in the process of assimilation. He planted cuttings of the same size in pots of the same soil. One plant was allowed to grow freely. The second was covered with a globe of Euphos glass which absorbs the ultraviolet radiation completely and partially absorbs blue and violet rays. The third plant was covered with a globe of clear glass which absorbs some of the ultraviolet rays of shortest wave-length in solar radiation. The globes were arranged so that they were ventilated. These ex- periments were repeated upon many plants for many years. The plant under Euphos glass grew much larger than the others and, although green, it reminded him of an etiolated plant. The plant under ordinary clear glass EFFECTS UPON LIVING MATTER 221 grew larger than that which grew freely but not as large as the one from which ultraviolet radiation was excluded. Schanz concluded that the form of plants is altered by radiation of short wave-lengths and most of all by ultra- violet radiation. He also concluded that, owing to the greater absorption of the radiation of the shorter wave- lengths by the plant material, these rays are less effective. Schanz claims that plants brought from high altitudes to low ones act as though, in the latter environment, they were protected by Euphos glass and that this is due to the relatively less amounts of ultraviolet energy at the lower altitudes than that to which the plants were ac- customed at high altitudes. Of course, there are many other differences such as temperature and humidity but he feels that short-wave ultraviolet energy is an important factor. The differences in the extent of the solar spec- trum are not great for various altitudes as seen in Chapter II, however, there are perhaps greater differences in rela- tive total amounts. Many other experiments indicate that there is much of interest to be learned regarding the role that ultraviolet radiation plays in plant life. References 1. Proc. Amer. Acad. Arts and Sci. 51, 1916, 640. 2. Klin. Monatsbl. Augenh. 1909, 721. 3. J. Amer. Med. Assn. Dec. 10, 1910. 4. Arch. f. Ophth. 69, 1908, i. 5. Proc. Roy. Soc. 85, 1912, 319. 6. Arch. f. Ophth. 71, 1909, 573. 7. Mitteil. Augenkl. zu Stockholm, Jena, 1898, 31. 8. Comp. Rend. 1883, 509. 9. J. Roentgen Soc. Eng. 1918. 10. Comp. Rend. 158, 1914, 1509. 11. Proc. Amer. Acad. Arts and Sci. 51, 1916, 760. 12. Amer. J. Physiol. 34, 1914, 21. 222 ULTRAVIOLET RADIATION 13. Amer. J. Physiol. 39, 1916, 335. 14. Comp. Rend. 135, 1902, 315. 15. Proc. Roy. Soc. 1917, 33. 16. Amer. J. Trop. Dis and Prev. Med. 1915, 506. 17. Arch of Ophth. 44, 1915, 498. 18. Comp. Rend. 150, 1910, 52 and 549. 19. J. Frank. Inst. 1914, 681. - 20. Elec. World 62, 81. 21. Proc. A. I. E. E. 33, 1914, 1906. 22. U. S. Patent 1156947. 23. Comp. Rend. 159, 278. 24. Comp. Rend. 158, 1509. 25. Chem. Zentr. i, 1911, 1454. 26. Comp. Rend. Soc. Biol. 73, 323. 27. Chem. Zeit. 35, 1911, 806. 28. J. Frank. Inst. 1916, 264. 29. English patent 16110. 30. English patent 12333. 31. U. S. patent 1141056. 32. French patent 400921. 33. Science 1913, 24 and 374, 34. Chem. Abs. 1917, 1753. 35. Comp. Rend. 152, 1911, 1709. 36. Comp. Rend. 153, 1911, 979. 37. Comp. Rend. 156, 1913, 1858. 38. Comp. Rend. 158, 1914, 1575. 39. Zeit. Bio. 34, 1897, 5O5- 40. Zeit. Physiol. Chem. 83, 1913. 41. Chem. Abs. 1918, 486. 42. Tropical Light, 1916. 43. Light Energy, 1904. 44. Geographical Jour. May 1911. 45. Biol. Zentr. Berlin, 1919; Sci. Amer. Mo. Jan. 1920, 13; Pfluger's Arch. vol. 170. Plate XI. The white flame arc as used in dye-testing. CHAPTER XII VARIOUS PHOTOCHEMICAL EFFECTS A great amount of work has already been done in in- vestigating the applications of ultraviolet radiation in chemical research and industry but much remains to be done before the subject can be summed up in general- izations. As Sheppard * stated a few years ago, " We are only at the beginning of the conscious utilization of the powers of light, as distinct from the unconscious enjoy- ment of them." In chemistry, radiation has many pos- sibilities because it is an uncontaminating catalytic agent and it throws light on chemical constitution. The ultraviolet radiations reaching the earth from the sun lie between about 290 and 400mji as seen in Chapter II. In other words, it is almost totally confined to the near ultraviolet region. No sweeping statements of the chemical effects can be made, for it is certain that there would be many ex- ceptions; however, by way of introduction a few will be harzarded. The near ultraviolet radiations, 300 to 400m^, are not particularly effective in destroying micro-organ- isms but their great intensity in sunlight tends to make up this deficiency to some extent. They act somewhat similar, chemically, to the violet and blue rays, producing exothermic changes perhaps as a rule. The middle ultraviolet radiations, 200 to SOOmjA, are readily produced by the arc, spark, and quartz mercury arc. They exert powerful germicidal action, coagulate albumen, and perform chemically in a manner similar to many enzymes and catalyzers in producing decomposition. They are readily absorbed by many substances and there- 223 224 ULTRAVIOLET RADIATION fore are responsible for many chemical reactions. Many endothermic reactions are prolonged at their expense and they also are responsible for various exothermic reactions of an irreversible character. The extreme ultraviolet radiations, 100 to 200mpi, pos- sess similar bactericidal action. Chemically, they often act synthetically, producing reversible reactions of an endothermic nature. The chemical properties of radia- tions of shorter wave-length than 150m[x are little known. They are absorbed very readily by even thin layers of gas. The chemical applications cannot be treated extensively in a single chapter but it is hoped that this account con- tains useful material which is necessary from the view- point of this book. The work of many investigators has been abstracted but there still remain many other works not alluded to here. Ellis and Wells 2 have published an extensive series of abstracts to which the author is in- debted for some references otherwise out of convenient reach. In considering the results of exposure to ultraviolet radiation it is necessary to distinguish between purely chemical action, the production of nuclei, and the forma- tion of carriers of electricity. For example, it is likely that the transformation of oxygen into ozone may be con- sidered a purely chemical action quite independent of the formation of nuclei or ions. The production of nuclei is perhaps the result of chemical action. Apparently the production of ions is due to ultraviolet radiation which is selectively absorbed. The radiation of short wave-lengths is chiefly responsible for the production of ions. Saltmarsh 3 found that nuclei, produced in air by ultra- violet radiation, were not affected by an electric field after they had travelled through a few centimeters of air. The nuclei are equally effective in condensing water, toluene, and turpentine vapors. No nuclei were formed by ultra- violet radiation in the absence of oxygen and carbon VARIOUS PHOTOCHEMICAL EFFECTS 225 dioxide. Oxygen containing ozone has nuclei for con- densation which act the same as those provided by ultra- violet radiation. According to Saltmarsh, it seems prob- able that the nuclei formed by ultraviolet radiation do not cause condensation by chemical action. They probably act like dust particles as centers on which condensation is possible. The nuclei formed by the ultraviolet radiation from the sun have figured prominently in theories concerning rainfall. Certainly they play a part as condensation nuclei but how important a part is a matter of conjecture. It is interesting to compute the number of negative electrons brought to earth in a certain amount of rainfall. The potential differences between a cloud and the earth arrived at in this manner are great enough to account for light- ning discharges. Of course, there are various assump- tions involved but the computations are interesting. Berthelot and Gandechon 4 have investigated a variety of photochemical effects of ultraviolet radiation. They found that cyanogen is oxidized to carbon dioxide and nitrogen ; hydrogen is not appreciably oxidized by oxygen at ordinary temperatures; formic acid to some extent is formed from mixtures of acetylene and oxygen and also of ethylene and oxygen. Acetylene, cyanogen, ethylene, and oxygen decrease in volume due to polymerization. Acetylene is transformed into a yellowish solid ; cyanogen into paracyanogen ; ethylene yields a liquid polymer; and oxygen is partially converted into ozone. Air and mix- tures of oxygen and nitrogen are not appreciably affected. Nitrous oxide alone or mixed with oxygen yields nitric acid, and sulphur dioxide alone or mixed with oxygen yields sulphur and sulphuric acid. The same investigators 5 found that solutions of am- monia in the presence of air or of oxygen are oxidized to nitrites but no nitrates are formed. Ammonium salts were oxidized to nitrites and urea is transformed to am- 226 ULTRAVIOLET RADIATION monia and then into the nitrite. Various other organic nitrogen compounds act similarly. They also found that ultraviolet radiation is capable of converting nitrates into nitrites and of liberating nitrogen by decomposing a solu- tion of ammonium nitrite. Under the action of ultraviolet radiation allylene is converted to a white solid and methane does not polym- erize, but in the presence of oxygen it suffers the loss of hydrogen resulting in higher homologues of the paraffin series. Ketoses, in the solid state or in solution, are de- composed by ultraviolet radiation. Erythrulose, sorbose, laevulose, and perseulose decompose somewhat more slowly than dihydroxyacetone. Carbon monoxide is evolved in these cases, the reaction consisting of the elimination of the carbonyl group. The same investigators 6 found that carbon monoxide and oxygen, under the radiation from the mercury arc, yield some carbon dioxide and the latter with hydrogen yields water and formaldehyde. Carbon dioxide in the presence of phosphorus yields carbon monoxide under the influence of ultraviolet radiation from the quartz mercury arc. Water is produced by the action of ultra- violet radiation upon a mixture of oxygen and hydrogen. Formaldehyde is formed from hydrogen and carbon monoxide; it is decomposed in the presence of nitrogen by long exposure to the radiation; and it is also formed by direct action of the radiation upon carbon monoxide and gaseous ammonia. Berthelot and Gandechon 7 studied the influence of ultraviolet radiation of various wave-lengths in a number of chemical reactions. For example, equal volumes of carbon monoxide and ammonia unite to form formamide on a few hours' exposure to the radiation from a quartz mercury arc of wave-lengths shorter than 200m^. The action is slower for the middle ultraviolet but is not pro- duced by the near ultraviolet, 300 to 400mjx. Hydrogen VARIOUS PHOTOCHEMICAL EFFECTS 227 chloride 8 is decomposed only by radiation of shorter wave-length than 200m^. Hydrogen iodide is decom- posed by blue and violet radiation. Hydrogen bromide is decomposed rapidly by radiation shorter than 200mji in wave-length. The same investigators studied water- vapor, hydrogen sulphide, telluride and selenide and carbon and nitrogen groups most of which exhibit decom- position on exposure to ultraviolet radiation. Gerretson 9 has considered the effect of radiant energy upon organic compounds and concludes that it depends upon wave-length. In general, ultraviolet radiation de- composes and radiation of longer wave-length polymerizes and condenses. Ultraviolet radiation decomposes carbon dioxide into the monoxide and water. It generally polym- erizes hydrocarbons; for example, acetylene yields benzene and a resinous compound. It decomposes alco- hols, aldehydes, and ketones. Esters yield carbon mon- oxide, carbon dioxide, hydrogen and a hydrocarbon. Ethers decompose similarly to alcohols. Acids evolve carbon dioxide, a hydrocarbon, and some carbon mon- oxide. Dibasic acids lose carbon dioxide leaving a mono- basic acid. Bioses yield monoses ; monoses evolve carbon monoxide, methane, and hydrogen. The complex absorp- tion by aldehydes and ketones resolve into elementary absorptions. The solvent is influential in the photo- chemical reaction. Sugars can be produced from formal- dehyde. Photo-chemical polymerization, isomerization, and condensation are common. Photochemical hydrolysis occurs; for example, most esters are saponified by water under exposure to radiation. Tian 10 found that water on exposure to ultraviolet radiation is decomposed into hydrogen and hydrogen peroxide and the latter is decomposed with the result that oxygen is liberated. Thus hydrogen and oxygen are evolved from water exposed to ultraviolet radiation. It is claimed that the effective radiation is in the vicinity of 228 ULTRAVIOLET RADIATION 185mpi. The quartz mercury arc emits several lines in this neighborhood. The aluminum spark emits several lines of approximately this wave-length and its radiation decomposes water in the same manner. It has been found by others that solar radiation decomposes water into hydrogen and hydrogen peroxide but it contains no energy of shorter wave-length than about 290m|i. Schopper 11 claims that solutions of coumarin such as a one per cent solution of dimethylamino-methylcoumarin absorbs ultraviolet radiation and may be used to protect the fabric of balloons from deterioration by solar radia- tion. Flour has been treated with oxidizing agents such as peroxides and then exposed to ultraviolet radiation to decompose the peroxide. Tian 12 found that the reaction between oxygen and hydrogen peroxide is increased by ultraviolet radiation. He claims that radiation between 250 and SOOmpi readily decomposes hydrogen peroxide but the decomposition of pure water is not accomplished excepting by radiation of extremely short wave-lengths. If the water con- tains oxygen, the latter combines with the freed hydrogen thus increasing the formation of hydro- gen peroxide. The oxygen is also transformed to some extent into ozone which reacts with the hydrogen peroxide. He found the conditions favorable to the formation of hydrogen peroxide are, powerful ultra- violet radiation of extremely short wave-length, the exposure of the water in thin layers, and the elimination of the influences and conditions which tend toward the decomposition of hydrogen peroxide. This decomposi- tion is diminished by using pure water and by eliminating the ultraviolet radiation of moderate wave-lengths. He recommends the use of low-voltage new quartz mercury arcs. Jaubert 13 patented a horizontal mercury arc surrounded VARIOUS PHOTOCHEMICAL EFFECTS 229 by a jacket in which liquids containing air or oxygen were circulated for sterilization. The ultraviolet radiation produces ozone from the oxygen and the ozone sterilizes the liquid. Thiele 14 investigated the influences of ultraviolet radia- tion from a quartz mercury arc upon many chemical re- actions. It was found that the combination of hydrogen and oxygen, with and without the presence of water- vapor, was hastened. Hydrobromic acid was decomposed by air. Carbon monoxide and oxygen also combined. Hydrochloric acid and air interacted. Hydrogen peroxide was formed from oxygen and water to some extent. Potassium nitrate was decomposed into the nitrite and oxygen. Sulphuric anhydride was formed on passing air into water containing sulphur in suspension after the water and sulphur had been exposed to ultraviolet radia- tion for several hours. The decomposition of potassium percarbonate, ammonium persulphate, hydrogen peroxide, formic acid, ammonium oxalate, egg albumen, and pep- tone was accelerated. Jacquier 15 has patented a process of precipitating metals from solutions of their salts by the use of ultraviolet radiation in the presence of aluminum. He suggests the use of mercury arcs between aluminum plates in a recep- tacle containing the solution of the metallic salt. Winther and Howe 16 have studied the effect of ultra- violet radiation on the photo-chemical decomposition of iron oxalate, acetate, succinate, tartrate, and citrate. Matthews and Elder have patented a process of pro- ducing compounds of sulphur dioxide and unsaturated hydrocarbons by the mixture of liquids exposed to ultra- violet radiation. The butylene product is a clear white solid which renders celluloid less inflammable and can be used in varnishes and other transparent films. Stoklasa 14 surveyed the literature on sugar synthesis and also reported his own work. He found that formic 230 ULTRAVIOLET RADIATION acid is first produced by ultraviolet radiation acting upon formaldehyde in the presence of caustic potash and air or oxygen. The formic acid is then decomposed into carbon dioxide and water. Lenard and Wolff 18 noted that metals are disintegrated when exposed to ultraviolet light and Svedberg 17 by applying this knowledge has produced colloidal solutions of metals. He placed the clean metal in a dish contain- ing a dispersion oxide medium and allowed the radiation from a quartz mercury arc to fall upon it. In the course of a few minutes' exposure he obtained colloidal solutions of lead, copper, tin, and silver, but was unsuccessful in the case of cadmium, platinum, and aluminum. By ex- perimenting with silver and lead in water, ether, ethyl and isobutyl alcohol, acetone, ethyl and amyl acetate, he found that the dispersion medium was an important factor in the result obtained. The colloidal particles obtained in this manner are said to be very small. Svedberg 20 found that the effect is decreased by oxidation of the metallic surfaces and that the active rays were those of shorter wave-length than 410m|i. He obtained interest- ing data on colloidal gold. Nordenson 21 proceeded in a manner similar to Svedberg in investigating the production of colloidal metals. He immersed the metals in water, alcohol, and other liquids and exposed them to ultraviolet radiation, X-rays, and even to the radiations from radium. All these radiations were effective but ultraviolet was more powerful than the others. He obtained colloidal solutions from silver, copper, mercury, lead, tin, and zinc, but not from gold, iron, nickel, manganese, platinum, cobalt, chromium, and bismuth. On immersing silver in pyridene or benzene no colloid was obtained but in ethyl alcohol, acetone, water, and methyl alcohol, colloid metal was formed. He concluded that the hydrogen peroxide formed by exposure of certain solutions to the ultraviolet radiation attacks VARIOUS PHOTOCHEMICAL EFFECTS 231 the metal, and the resulting oxide or hydrate is dissolved and reduced to metallic colloid by the action of the radi- ation. The radiation from a quartz mercury arc colors phenol, in the presence of oxygen, very quickly to a red. Acetylene and methane unite when heated in the pres- ence of various metals. It is said that ultraviolet radia- tion produces the same result as heat in this case. Flour is slowly bleached by ultraviolet radiation. Chlorine gas, exposed to the silent electrical discharge in a manner similar to that used in the production of ozone, is transformed into a more active form. Zinc ethyl is quickly decomposed by the radiation from the quartz mercury arc. Carbonyl chloride is slowly decomposed by ultraviolet radiation of short wave-lengths. Cyanogen polymerizes when exposed to solar radiation. On exposure to ultraviolet radiation in the presence of oxygen it gives para-cyanogen and nitrogen. Acetone is not decomposed by solar radiation but is decomposed into carbon monoxide and ethane by the radiation from the quartz mercury arc. Formamide slowly decomposes on exposure to ultra- violet radiation. The blue, violet, and ultraviolet radiations are respon- sible for the transformation of yellow phosphorus which is soluble in carbon bisulphide into red phosphorus which is insoluble. Weigert and Bohm 22 investigated the effect of ultra- violet radiation on the decomposition of ozone and the formation of water by exposing mixtures of hydrogen and ozonized oxygen to the radiation from a quartz mercury lamp. According to Mack 23 chlorate may be oxidized to per- chlorate by means of oxygen exposed to ultraviolet radia- tion. 232 ULTRAVIOLET RADIATION By exposing chlorine and sulphur dioxide to the radia- tion from a quartz mercury arc, oxychloride is formed even in glass vessels. The electrical conductivity of amorphous sulphur is in- creased by exposure to radiation, that of shorter wave- length than 280ml! being the more active. Boeseken and Cohen 24 have investigated the photo- chemical reduction of ketones by anhydrous alcohol under the influence of solar radiation and of quartz mercury-arc radiation. The particularly active radiation appears to be in the region of 410mji. Behthelot and Gaudechon 25 found that ultraviolet radia- tion photolyzes di-saccharides. Hydrolysis first takes place without the evolution of gas, then later the decom- position of the monoses, produced at first, results in evolu- tion of gas. During the photolysis of sucrose to dex- trose and levulose, and of maltose to two molecules of dextrose, the solutions remain neutral. The exposure of levulose, dextrose, maltose, and sucrose to the radiation from the quartz mercury lamp results in the evolution of carbon monoxide, methane, hydrogen, and carbon dioxide. They 26 also found that acetaldehyde was decomposed by solar radiation but acetic acid and ethyl alcohol were unaffected. However, the radiation from a quartz mer- cury arc photolyzed each of them. Acetaldehyde is de- composed to carbon monoxide and methane but is also polymerized to paraldehyde, the latter being decomposed into carbon monoxide, ethane, and some complex com- pounds. Ethyl alcohol is decomposed into acetaldehyde and hydrogen, but there is also a further decomposition of the former into carbon monoxide and ethane. Acetic acid decomposes into carbon monoxide, carbon dioxide, and other gases. The same investigators found ammonia and also hydro- gen phosphide were decomposed by the radiation from the quartz mercury arc. Hydrogen arsenide in a quartz VARIOUS PHOTOCHEMICAL EFFECTS 233 tube suffered decomposition but was unaffected in a glass tube. This was also true of carbonyl chloride. Sulphur fluoride and methane respectively were unaffected even in quartz tubes. Zinc ethyl yielded zinc and ethane free from ethylene. They 27 claim that the photo-chemical effect of ultra- violet radiation is to lower the temperature of reaction. For example, decompositions and oxidations are caused by it below 100 C. which would not be consummated ordinarily at much higher temperatures. The reactions produced are reversible and the radiation possesses great polymerizing power especially in some cases. In several respects the photo-chemical action is similar to the re- actions produced in plants. Methyl compounds are easily synthesized but the production of ethyl compounds is less certain and more difficult. Ultraviolet radiation greatly accelerates the chlori- nation of toluol. Some results indicate that visible radia- tion retards it. Baskerville and Riederer 28 found that ultraviolet radiation did not materially accelerate the reaction be- tween natural gas and chlorine. Mott and Bedford 29 passed chlorine gas into benzol, isoamyl, chloride, and isopentane and investigated the effectiveness of the white flame arc and the quartz mer- cury arc in chlorination. By using quartz and glass con- tainers they showed that the white flame arc radiated energy of more desirable wave-lengths in general than the quartz mercury arc for the chlorination reaction. On using a glass vessel the time necessary for the free chlo- rine to disappear was not materially longer than in the case of a quartz vessel when both were exposed to the radiation from the white flame arc. However, for the quartz mercury arc the time for chlorination in a glass vessel was much greater than that for a quartz vessel. They claim that the white flame arc is superior to the quartz 234 ULTRAVIOLET RADIATION mercury arc for this purpose. They summarize the sen- sitiveness of halogens and their compounds somewhat as follows: Flourine in its compounds is wonderfully stable to radiation. Silver fluoride is not sensitive like the other halogen salts of silver. Sulphur fluoride does not yield to ultraviolet (or visible) radiation although the sulphur oxides and hydride are sensitive to ultraviolet radiation. The great stability of calcium fluoride is associated with a remarkable transparency to ultraviolet radiation. Chlorine responds chiefly to blue, violet, and ultraviolet and its absorption spectrum extends further into the ultraviolet than that of bromine. The photo-sensitiveness of silver chloride is more marked in the ultraviolet than silver bromide. Bromine in the free form responds to radiation less in the ultraviolet than chlorine. On the other hand, hydrobromic acid is more easily decomposed by ultraviolet radiation than is hydrochloric acid. Iodine extends its spectrum still further towards the long wave- lengths and hydriodic acid is decomposed by ordinary blue and violet light. Bordier 30 studied the effect of ultraviolet radiation on iodine and starch iodide. He states that the action is perhaps to remove the charges from the colloidal particles of these substances with the result that the iodine in the aqueous and alcoholic solution combines with the hydro- gen ion to form hydriodic acid. Pougnet 31 investigated the influence of the radiation from a quartz mercury arc on a solution of mercuric chloride and on certain mercury salts. Mercuric chloride in water is gradually altered into mercurous chloride and mercury. Most of the salts examined were darkened whether in a dry or moist state. Keyes 32 has described an apparatus for the production of chlorine and bromine derivatives of hydrocarbons. It consists essentially of a quartz mercury arc with a straight tube which is surrounded by a quartz jacket which may VARIOUS PHOTOCHEMICAL EFFECTS 235 also be exhausted. Another jacket surrounds this and the halogen mixed with the hydrocarbon is admitted to this outer jacket. The vacuum jacket protects the liquids in the outer jacket from the heat. The apparatus may be operated continuously. Kailan 33 found that strontium iodide in neutral solu- tions decomposes faster than barium iodide, and the latter more rapidly than potassium or magnesium iodide, but when exposed to ultraviolet radiation, barium iodide decomposes more rapidly than strontium iodide although both of them still decompose faster than the other two. In their experiments they found the effect of ultraviolet radiation and penetrating radium rays to be the same, but under the conditions of their experiments the decom- position produced by the former was several hundred times faster than by the radium rays. Marmier 34 decomposed solutions of sodium thiosul- phate by ultraviolet radiation, sulphur and hydrosulphite being formed. The latter is also decomposed by con- tinued exposure, the liquid containing chiefly sulphite. When the solution contained more than 5 grams per liter, no hydrosulphite was formed. Sodium sulphite solutions are not oxidized by ultra- violet radiation in the absence of air but when air is present this radiation hastens the oxidization. Solutions of oxalic acid are decomposed slowly by ultraviolet radiation but the addition of uranium salts hastens decom- position. Potassium permanganate and potassium bi- chromate are quite unaffected by ultraviolet radiation. Matthews, Bliss, and Elder 35 obtained a patent for re- moving the halogen hydride from the compounds contain- ing halogen by exposing the hydride vapor to ultraviolet radiation. McLennan 36 experimented with iodine crystals in an evacuated quartz tube which was inserted in a mercury vapor tube. He found that the fluorescence of the iodine 236 ULTRAVIOLET RADIATION vapor is excited by radiation between 180 and 210m[x. Resonance spectra could not be obtained with iodine vapor with radiation shorter in wave-length than the green line, 546mpi. The fluorescence of iodine was obtained at tem- peratures from that of the room up to 1000 C. Mercury iodide and potassium iodide exhibited fluorescence spec- tra of their own at a temperature above 326 C. when excited by the radiation from the mercury arc. A German company developed a process for producing iodine derivatives of the paraffine series in the form of vapor by exposing to ultraviolet radiation mixtures of hydrocarbons and hydriodic acid. Cohen 38 found hydrochloric acid to be decomposed by the ultraviolet radiation from a quartz mercury arc. The result at a comparatively low temperature was as great as ordinarily obtained at 1500 C. Fischer and Braehmer 39 devised a double-walled mer- cury arc, the quartz walls being separated from each other by about 1 mm. Electrolytically produced oxygen was passed through the annulus and thus was exposed to intense ultraviolet radiation. The yield of ozone was increased very much by cooling the annulus with water. Ozone is a common product of ultraviolet radiations of the shorter wave-lengths emitted by the quartz mercury arc and some other sources so that its presence often must be reckoned with. According to Pribram and Franke 40 the radiation from a quartz mercury arc produces no visible change upon a redistilled 30 per cent solution of formaldehyde, but if the liquid is distilled after exposure a white residue remains. The distillate and the residue are strongly reducing. Kailan 41 found dilute solutions of tartaric, succine, malonic, acetic, and oxalic acid were slightly decomposed by the radiation from a quartz mercury arc. No de- composition took place when the solutions were in glass vessels. The decomposition of a dibasic acid was has- tened by introducing an alcoholic hydroxyl group. Plate XII. The carbon arc as used in blue-printing. VARIOUS PHOTOCHEMICAL EFFECTS 237 Sulphuric acid 42 is produced by exposing a mixture of sulphur dioxide and air (or oxygen) to ultraviolet radia- tion when the gases are at a temperature above 300 C. Cohen and Becker 43 investigated the production of sul- phur trioxide by exposing sulphur dioxide and oxygen to the radiation from a quartz mercury arc. No oxides of nitrogen were formed when air was used instead of oxygen. Lesure 44 found the radiation from a Cooper-Hewitt lamp to exert influences as follows: Solutions of silver nitrate, eserine salicylate, apomorphine hydrochloride, arbutin, and guaiacol were slightly discolored in about 30 minutes. Solutions of cocaine hydrochloride, mercury benzoate and bichloride, sodium cacodylate, calcium gly- cerophosphate, quinine bichloride, pilocarpine hydro- chloride, and artificial serums were unaffected by ex- posures of thirty minutes. Olive oil was bleached. Cul- tures of bacillus coli were added to solutions of aucubin and of gentiopicrin. These solutions were sterilized in 30 seconds to 30 minutes. Bierry, Henri, and Ranc 45 found that a 5 per cent solu- tion of sucrose is appreciably hydrolyzed on exposure to the radiation from a quartz mercury arc for 20 hours. After 40 hours' exposure the solution contained formal- dehyde. In the absence of calcium carbonate a gas is liberated which consists largely of carbon monoxide. Holbronner and Bernstein 46 found that ultraviolet radi- ation effected a permanent vulcanization upon rubber solutions containing sulphur. They made various experi- ments on layers of rubber solutions carried underneath a quartz mercury arc provided with a cooling jacket. It is stated that the vulcanized rubber forms a stable gel and is not precipitated after several months or by pro- longed heating at 80 C. Bernstein 47 obtained a patent for vulcanizing India rubber solutions containing sulphur by exposing them to the radiation of a quartz mercury 238 ULTRAVIOLET RADIATION lamp. A solution of rubber in benzene containing 6 per cent of rubber and 0.6 per cent of sulphur is vulcanized in less than one minute when in a thin layer about 0.5 mm. thick. A xylene solution of rubber containing 6 per cent of sulphur was evaporated on a quartz plate and then exposed to a quartz mercury arc at a distance of 15 cm. After 40 minutes' exposure on both sides the film exhibited the properties of vulcanized rubber and was found to contain 2.56 per cent of combined sulphur. 48 The percentage of combined sulphur increases with the duration of exposure to the ultraviolet radiation. Ap- parently excessive exposure tends to deteriorate vulcan- ized rubber. 49 It has been claimed that wine, after passing the fer- mentation stage, on being exposed to ultraviolet radiation in thin films, changes in taste and color and attains the characteristics of old wine. It is said that the duration of treatment is not long and that this application made some advances in France before the war. Red wines are bleached to some extent by ultraviolet radiation. In fact, many of the fruit colors are rapidly faded by ultra- violet radiation. Henri and Ranc 50 found that a 10 per cent aqueous solu- tion of glycerine exposed for several hours to the radiation from a quartz mercury arc resulted in the decomposition of the glycerol molecule and the production of formal- dehyde and acids and also other aldehydes. Others have obtained the same results. Apparently the presence of hydrogen peroxide increases the decomposition. The same investigators with Bierry 51 experimented with glycerol and ultraviolet radiation from quartz mer- cury arcs. They used various concentrations of glycerol from 1 to 100 per cent and varied the temperature and admitted and excluded air. The action of ultraviolet radiation on egg and serum proteins has been studied by several investigators. (See VARIOUS PHOTOCHEMICAL EFFECTS 239 Chapter XI). In general, it is the radiation shorter than SOOmpi in wave-length which is effective in coagulating them. Schanz 52 concluded that the change produced by ultraviolet radiation could be regarded as a gelatinization of the protein. In acid solutions the precipitation pro- duced by ammonium sulphate and sodium chloride was increased but in alkaline solutions it was decreased. Henri 53 determined the absorption of egg albumen and found it to be negligible at SOOmjj, but rapidly increased with decreasing wave-length. The reaction of chlorine and hydrogen is purely photo- chemical. Bunsen and Roscoe ascribed the maximum activity to radiation between 395 and 413m| J i. They used a gas bulb in a thermostat with glass walls and obtained their radiant energy from a mercury arc. They studied the temperature coefficient of the reaction and found it to be small for ultraviolet radiation, greater for violet and blue, and greatest for green light. It was intermediate for total (white) light. The temperature coefficient also has been investigated by Padoa and Butironi. 54 Favre and Silbermann 55 have verified the variation of catalytic action with wave-length. Ekely and Banta 56 found that lead phthalate on ex- posure to the radiation from a quartz mercury arc de- composed and a yellowish brown substance was formed. Mercuric phthalate exposed in the same manner was not decomposed. According to de Fazi " ultraviolet radiation favors alcoholic fermentation and beer yeast exposed to it in- creases in activity. A 14-hour exposure to the radiation from a quartz mercury arc about 20 cm. distant did not destroy the activity of the yeast. Davis 58 has patented apparatus which exposes air to violet and ultraviolet radiations and partially ozonizes it. Iron and carbon plates are connected to an electric circuit, preferably alternating, and air is drawn between these 240 ULTRAVIOLET RADIATION plates. The carbon is supported on a glass plate and a thin silica plate is supported on an iron plate. The temperature effect produced by the quartz mercury arc in chemical applications is sometimes confusing. Weigert 59 has attempted to avoid this by water-cooling in such a manner that the reaction proceeds outside the lamp itself, but the reaction receives radiation which has passed through non-absorbing media. Cohen and Becker 60 studied the decomposition of car- bonyl chloride into carbon monoxide and chlorine under the influence of the radiation from the quartz mercury lamp. They found that the effective radiation was of shorter wave-length than 265mpi. Cassel 61 studied the influence of radiation from an arc lamp upon mixtures of alcohol with chloracetic and bro- macetic acid respectively. He found the reaction was not sensitive to radiation of longer wave-lengths than 250mjA. The decomposition under the influence of the short-wave ultraviolet resulted in the formation of methyl alcohol, acetic aldehyde, and the halogen acid. The chlorine compound is less easily decomposed by alcohol than the bromine compound. Various suggestions have been made regarding the treatment of seeds with ultraviolet radiation but con- clusive data pertaining to economic advantages is lacking. Pougnet 62 found that green vanilla pods under exposure to ultraviolet radiation gave forth the odor of vanillan. He 63 also exposed other plants to the radiation from a quartz mercury arc. Those containing coumarin emitted the odor of coumarin, and other plants yielded character- istic odors. Plants grown under glass appear to be more sensitive to ultraviolet radiation than those grown outdoors. An alcoholic solution of chlorophyll is not readily decomposed by ultraviolet radiation, but this radiation appears to de- velop the green coloring matter in some leaves more rapidly than solar radiation. VARIOUS PHOTOCHEMICAL EFFECTS 241 Euler 64 found that aqueous solutions of haloacetic acids exposed to ultraviolet radiation from the mercury arc resulted in decomposition at 18 C. equal to that at 100 C. without exposure to ultraviolet. Ultraviolet radiation of short wave-lengths causes the evolution of oxygen from potassium nitrate. It has been reported that ultraviolet radiation causes polymerization of vinyl esters and the solids yielded are substitutes for celluloid. These solids are odorless and non-inflammable. The polymerization of vinyl acetate and of vinyl chloracetate is aided by organic peroxides. Pougnet 65 investigated the action of the radiation from a mercury arc on a 5 per cent solution of mercuric chlo- ride. At a distance of 15 cm. from a 110-volt, 4-ampere lamp the solution at once became clouded by the calomel which was formed. Most of the mercury salts were af- fected by ultraviolet radiation. Ross 66 has studied the effect of the radiation from the aluminum spark upon solutions of potassium iodide and upon the reduction of chlorates and bromates. A number of patents have been obtained which cover certain chlorination processes. One 67 describes the use of a mercury vapor lamp or large tungsten lamp for chlorinating liquid hydrocarbons such as gasoline, petro- leum oils, toluol, and benzol. Another 68 pertains to the production of unsaturated hydrocarbons from petroleum oils. Others 69 involve the chlorination of toluol and the production 70 of benzol chloride, benzyl chloride, and ben- zotrichloride. Another 71 pertains to the production of tetrachlorotoluene by passing dry chlorine into toluene in the presence of iron. A patent 72 involves the bromination of hydrocarbons by the use of a mercury arc. Bedford 73 exposed a mixture of chlorine and natural gas, in a vessel containing ice, to the radiation from a flame arc. Besides water from the melted ice he ob- tained a heavy liquid consisting of methylene chloride, 242 ULTRAVIOLET RADIATION chloroform, carbon tetrachloride, and chloroethanes. These were also dissolved in the water to some extent. Tolloczko 74 describes the chlorination of natural gas with aid of the mercury arc. Phosgene has been produced by uniting carbon dioxide and chlorine under the influence of ultraviolet radiation. Cohen and Sieper 75 found that the action of ultraviolet radiation of shorter wave-length than 254m( J i on carbon dioxide results in decomposition which is .greatly de- creased by the presence of water-vapor. They found that no formic acid or formaldehyde is produced but a mixture of equal volumes of hydogen and carbon dioxide results in the formation of formaldehyde. They found that the presence of carbon monoxide or dioxide does not increase the amount of ozone formed on exposing oxygen to ultra- violet radiation. The ultraviolet radiation in the middle spectral region is used on a large scale for bleaching olive oil. Linseed oil, other oils, and resins are bleached by ultraviolet radi- ation. Rohm 76 has prepared a substitute for drying oil in paints which consists of a solution of polymerized acrylic acid ester in acetone or other solvents. When this solu- tion is exposed to ultraviolet radiation it is transformed into a colorless, transparent, elastic substance soluble in the usual oil-solvents. Klatte and Rollett 77 have patented a process for making celluloid substitutes by polymerizing vinyl esters under the influence of the mercury arc and other sources. Mott 78 made an extensive investigation of the use of the flame arc in dye-testing and concluded that the high- amperage arc is more powerful than sunlight. He ob- tained essentially similar results in fading dyes as he obtained with June sunlight and at a much greater speed. The best June sunlight for 50 hours produced an effect equal to that produced by a 28-ampere white flame arc in VARIOUS PHOTOCHEMICAL EFFECTS 243 10 to 20 hours at a distance of 10 inches. The blue flame arc produced results with some dyes unlike that of day- light. This arc is especially rich in ultraviolet radiation. Some dyed clothes faded in an hour under exposure to the radiation from the white flame arc. Lithopone showed maximum darkening at low temperatures. Mott pre- sented a brief summary of the literature. The author found that lithopone is darkened by radi- ation from the quartz mercury arc chiefly between 300 and 350mjj, in wave-length. By focusing the image of a quartz mercury arc upon a lithopone surface darkening was produced in a few seconds. Both the commercial magnetite arc and an experimental iron arc produced blackening in a few minutes. Hydrogen peroxide black- ened the lithopone at once. O'Brien 79 made an extensive study of lithopone using a flame arc. Scheurer 80 tested benzo colors by exposing them to radiation from the sun and to the quartz mercury arc. Under the former they were only slightly faded but under the quartz mercury arc they were quite faded in 24 hours. Indigo was less changed by the mercury radiation than by solar radiation. There are dye-testing equipments available commer- cially which employ various arcs. The testing of per- manency of colors is unsatisfactory when solar radiation is used owing to the variation of conditions. Chevreul showed in 1837 that many dyes are not faded in a vacuum but tumeric and prussian blue were excep- tions. Joffre 81 found many dyes to be unaffected in nitro- gen but picric acid was an exception. The action of air can be eliminated by the use of paraffin. According to Dufton 82 the radiations complementary to the color of the dye are most effective. This is merely a statement of the law of photo-chemical absorption but it is necessary to know what takes place in the ultraviolet region as well as in the visible. To state that a dyed color is not appre- 244 ULTRAVIOLET RADIATION ciably affected by light of the same color must not mis- lead one into forgetting that invisible radiation may be present. According to Brownlie, 83 ammonia, alcohol, or pyridine vapors enormously increase the action of radiation on dyes, and naphtha and chloroform slightly retard it. Toch 84 has discussed the influence of solar radiation on paints and varnishes. He found glass to be protec- tive. Varnish performs the useful function of excluding oxygen. He states that linseed oil bleaches due to the action of radiation on green chlorophyll. In dye-testing outdoors it is well to note that the at- mosphere in cities is usually more acid than in the country. In fact, it is said that the atmosphere in the country is alkaline. According to some investigators the most favorable condition for bleaching by solar radiation is one which is hot, moist, and alkaline. According to Gebhard 85 oxidation in the light may be quite different from that in darkness even when the same oxidizing agent is used. There appears to be no doubt that the fastness of dyes is decreased by moisture and that there is little or no action in a vacuum in a great many cases. Apparently the oxidation theory of fading has been confirmed by experiments and is adhered to by many, but the reduction theory is favored by some, at least in particular cases. Gebhard 86 has confirmed the oxidation theory and also states that permanency of dyeing affects permanency of the fiber. Oxidation and the formation of peroxide hy- drates are involved in the bleaching of many if not all dyes. Inks may be tested quickly by drawing lines on paper and exposing these to ultraviolet radiation. Ellis and Wells 87 have described a variety of experi- ments with oils, resins, and varnishes exposed to the radia- tion from a quartz mercury arc. Various effects of VARIOUS PHOTOCHEMICAL EFFECTS 245 bleaching and gelatinization were noted. Among the substances tested were cotton-seed, crude whale, linseed, and castor oils. Gray 88 has patented a process for producing fatty acids and esters from hydrocarbons with the aid of a mercury arc. Graul and Hanschke 89 have been granted a patent for a process for producing halogenated paraffin hydrocarbons by mixing chlorine with the hydrocarbon in the dark and then exposing to radiation from the mercury arc or other source. They describe the method of making chlor- hexane and heptane. Daree and Dyer 90 exposed clean bleached cotton cloth to the radiation from a mercury (glass tube) arc for a week. They concluded that the ultraviolet radiation con- verted cellulose into oxycellulose and it lost its tensile strength. Ellis and Wells 91 investigated the effect of ultraviolet radiation on the production of benzyl chloride, benzal chloride, and benzo trichloride by chlorinating toluol. They used the quartz mercury arc and passed an excess of toluol vapors mixed with chlorine through a quartz tube exposed to ultraviolet radiation. Boll 92 has described an investigation of the decom- position of the chloroplatinic acids. He found that they were influenced by radiation from the extreme yellow region to the extreme ultraviolet. 93 Luther and Forbes 94 investigating the oxidation of quinine by chromic acid under the influence of radiation found that the relative effects of the radiation from a mercury arc for 362mj > i and 406m|x were 75 and 100 re- spectively. When an aqueous solution of ferrous and mercuric chlorides is exposed to ultraviolet radiation, ferric chlo- ride and calomel are formed. Inasmuch as the reverse action takes place completely, Winther 95 proposed an 246 ULTRAVIOLET RADIATION electric battery based upon these reactions. The action is slow at ordinary temperatures but proceeds rapidly when the battery is short-circuited. He obtained as much as 0.1 volt and 1 milliampere. Kailan 96 has reported experiments upon water solu- tions of urea, benzoic, and formic acids. There appear to be many counteracting phenomena of different radiations such as ultraviolet, visible and infra- red. According to le Bon 97 infra-red radiation destroys chlorophyll and alters the color of tomatoes and arti- chokes. Pech 98 claims that the bleaching action of ultra- violet radiation on raw cotton is retarded slightly by visible radiation and greatly by infra-red. Ultraviolet radiation which produced erythema on an animal skin placed in a dark chamber in five minutes was found to produce the same effect in seven minutes when visible radiation was present and in ten minutes when infra-red was present. Many such counteracting influences may be suspected since Becquerel in 1872 found that certain samples of zinc sulphide phosphoresced under ultraviolet radiation but did not phosphoresce when infra-red was present. In this case visible radiation of shorter wave- length than yellow rays increases the phosphorescence. If a beam of infra-red radiation is projected upon a surface of zinc sulphide under excitation of near ultra- violet radiation a black spot appears amid the brilliantly phosphorescent surroundings. This black spot is at the place where the infra-red is incident, showing that the latter prevents the production of phosphorescence by the radiation of shorter wave-lengths. Curie " found that with a fluorescent substance no black spot or variation of intensity of the fluorescence could be observed at the place where the infra-red rays were con- centrated. T. Swensson 10 has described researches pertaining to the potential changes by ultraviolet radiation on oxidizing VARIOUS PHOTOCHEMICAL EFFECTS 247 agents. When mixtures of potassium dichromate and sulphuric acid are illuminated by a quartz mercury arc at 18, potential changes appear. The electrodes, bright or platinized platinum, need not themselves be illumi- nated. In the mixture of dichromate and sulphuric acid the potential rises, in the separate solutions of the com- ponents it falls, and it falls also when chromic acid is illuminated. Apparently the liberation of this chromic acid is not essential for the rapid first effect. References 1. Photo-chemistry, 1914. 2. Chem. Engr., Vols. 25-27. 3. Proc. Phys. Soc., London, 27, 1915, 357. 4. Comp. Rend., 150, 1910, 1169, 1327, 1517. 5. Comp. Rend., 152, 1911, 522. 6. Comp. Rend., 150, 1910, 1690. 7. Comp. Rend., 155, 1912, 207. 8. Comp. Rend., 156, 1913, 889, 1243. 9. Chem. Weekblad., 13, 1916, 220. 10. Comp. Rend., 152, ign, 1012 and 1483. 11. J. S. C. D., 1916, 356. 12. J. Chem. Soc., 108, 1915, 828. 13. French patent, 415-574. 14. Zeit. Aug. Chem., 22, 1909, 2472. 15. English patent, 17790. 16. Zeit. Wiss. Phot., 14, 1914, 196. 17. Zentr. Biochem. Biophys., 18, 370. 18. Ann. Phys. Chim., 37, 1889, 443. 19. Ber. 42, 1909, 4375. 20. Z. Chem. Ind. Kol., 6, 1910, 129 and 238. 21. Kolloid. Chem. Beihefte, 7, 1915, no. 22. Z. Phys. Chem., 90, 189. 23. J. Phys. Chem., 1917, 238. 24. Wetenschappen, 23, 1914, 765. 25. Comp. Rend., 155, 1912, 1016. 26. Comp. Rend., 156, 1913, 68 and 1243. 248 ULTRAVIOLET RADIATION 27. Comp. Rend., 151, 1910, 395. 28. J. Ind. Eng. Chem., 1913, 5. 29. J. Ind. Eng. Chem., 8, 1916, 1029. 30. Comp. Rend., 163, 1916, 205. 31. Comp. Rend., 161, 1915, 348. 32. U. S. patent, 1237652. 33. Chem. Abs., 1914, 14. 34. Comp. Rend., 154, 1912, 32. 35. English patent, 16828. 36. Proc. Roy. Soc., London, 91, 1914, 23. 37. German patent, 266119. 38. Ber., 42, 1909, 3183. 39. Ber., 38, 1905, 2633. 40. Ber., 44, 1911, 1035. 41. Mon. Chem., 34, 1913, 1209. 42. German patent, 217772. 43. Z. Phys. Chem., 70, 1910, 88. 44. Chem. Abs., 1911, 963. 45. Comp. Rend., 152, 1911, 1629. 46. Rubber Ind., 1914, 156. 47. English patent, 17195. 48. J. Frank. Inst., 1913, 345. 49. U. S. patent, 1240116. 50. Comp. Rend., 154, 1912, 1261. 51. Comp. Rend., 152, 1911, 535. 52. Arch. Ges. Physiol., 164, 1916, 445. 53. Comp. Rend., 135, 1902, 315. 54. Chem. Abs., 1917, 1356. 55. Ann. Phys. Chim., 37, 297. 56. J. Amer. Chem. Soc., 1917, 762. 57. Chem. Abs., 1916, 950. 58. U. S. patent, 1209132. 59. Z. Phys. Chem., 80, 1912, 67. 60. Ber., 43, 1910, 130. 61. Z. Phys. Chem., 92, 1916, 113. 62. Comp. Rend., 152, 1911, 1184. 63. Comp. Rend., 151, 1910, 355. 64. Ber., 49, 1916, 1366. 65. Comp. Rend., 161, 1915, 348. VARIOUS PHOTOCHEMICAL EFFECTS 249 66. J. Amer. Chem. Soc., 1906, 786. 67. U. S. patent, 1191916. 68. U. S. patent, 1220821. 69. U. S. Patent 1146142. 70. U. S. patent, 1202040. 71. English patent, 16317. 72. U. S. patent, 1198356. 73. J. Ind. Eng. Chem., 1916, 1090. 74. Chem. Abs., 1914, 1282. 75. J. Phys. Chem., 1916, 347. 76. German patent, 295340. 77. U. S. patent, 1241738. 78. Trans. Amer. Electrochem. Soc., 28, 1915, 371. 79. J. Phys. Chem. 19, 1915, 113. 80. Bui. Soc. Ind. Muhl., 80, 1911, 324. 81. Bui. Soc. Chim., Paris, i, 1889. 82. J. Soc. Dyers, 1885, 245. 83. J. Soc. Dyers, 1902, 295. 84. J. Soc. Chem. Ind., 37, 1908, 311. 85- J- Agew. Chem., 26, 1913, 79. 86. Farber Ztg., 21, 1911, 253. 87. Chem. Engr., 26, 1918, 114. 88. U. S. patent, 1158205. 89. U. S. patent. 1032822. 90. Chem. Abs., 1917, 1753. 91. Chem. Engr., 26, 1918, 182. 92. Ann. D. Phys., 2, 1914, 56. 93. Comp. Rend., 156, 1913, 138. 94. J. Chem. Soc., 1909, Abs. 96, 632. 95. Z. Electrochem., 18, 1912, 138. 96. Mon. Chem., 41, 1920, 305. 97. Comp. Rend. June 14, 1920, 1450. 98. Comp. Rend. May 25, 1920, 1246. 99. Comp. Rend., 172, 1921, 274. 100. Ark. Kem. Min. Geol. Stockholm, 7, 1917, i: Sci. Abs. No. 1113 (1920) and No. 1500 (1921). INDEX Abiotic action, 205 Absorption bands, 44 and frequency, 42 Acetaldehyde, 232 Acetate, 229 Acetic acid, 52, 236 Acetoacetates, 60 Acetone, 51, 52, 60, 62, 68, 231 Aceturic acid, 65 Acetylene, 225, 227, 231 flame, HI Acids, 62, 66, 227 Acrylic acid ester, 242 Actino-therapy, 217 Adularia, 74 Aesculine, 123 Agar, 210 Air, 30, 39 Albumen, 67 Alcohol, 55, 56, 66, 67, 69, 227, 244 ethyl, 51, 232 Aldehydes, 61, 66, 227 Alkaloids, 63, 186 Alloys, 95 silver and cadmium, 147 Allylene, 226 Alum, 74 Aluminum, 102 spark, 138, 141, 145 Amino-acids, 64 Ammonia, 44, 50, 52, 232, 244 Ammonium molybdate, 137 oxalate, 229 persulphate, 229 salt, 185, 225 sulphite, 66 Aniline, 65 Anilonoacetic acid, 65 Anthracene, 61, 121 Antimony, 102 spark, 143 Apoatropine, 63 Apomorphine hydrochloride, 237 Apophyllite, 74 Arbutin, 237 Arc, no, HI, 116, 134, !45> crater, 154 flame, 124, 126, 127 tungsten-mercury, 130 white flame, 125 Argon, 38, 135 Aristo arc, in Arragonite, 74 Atmosphere, 33 transmission of, 20 Atropine, 63 Bacillus coli, 237 communis, 67 Bactericidal action, 209, 224 Ballon fabric, 228 Barite, 74 Barium iodide, 325 spark, 140 Beer, 215 Beer's law, 47, 68 Beer yeast, 239 Benzal chloride, 245 Benzene, 43, 63, 66 derivatives of, 42 ethyl, 43, 66 halogen, 64 vapor, 41 Benzo colors, 243 Benzoic acid, 64, 246 Benzol, 51, 233, 241 chloride, 241 Benzotrichloride, 241, 245 251 252 INDEX Benzyl, 63 chloride, 241, 245 Beryl, 76 Bioses, 227 Bismuth, 99, 102 Bleaching, 215, 244 oils, 242 Blindness eclipse, 205 snow, 204 Blood, 213, 216 serum, 213 Bolometer, 166 Borax, 74 Boric oxide, 83 Brandes-Schiinemann's alloy, 95 Brashear alloy, 96 Bromacetic acid, 240 Bromination, 126, 241 Bromine, 41, 43, 55, 234 Butylene, 229 Cadmium, 102 arc, 140, 146, 147 spark, 141 vapor, 43 Cairngorm, 76 Calcite, 74 Calcium nitrate, 51 spark, 140 Calcspar, 75 Canada balsam, 52 Carbon, 102, 153 arc, in, 116, 124 bisulphide, 43 dioxide, 31, 39, 226, 242 monoxide, 38, 226, 229 tetrachloride, 242 Carbons, impregnated, 125 Carbonyl chloride, 231, 233, 240 Carborundum, 102 Catalyzer, 223 Cataract, 208 Catechol, 62 Celestite, 74 Cells, 215 Celluloid, 77 substitute, 242 Cellulose, 245 Cerium nitrate, 158 Chemical applications, 224 Chloracetic acid, 240 Chlorate, 231 Chloride, 233 Chlorination, 233, 239, 241 Chlorine, 41, 232, 234 gas, 231 Chlorites, 67 Chloroform, 242 Chlorophyll, 8, 9, 220 Chloroplatinic acids, 245 Chlorous acid, 67 Chromatic aberration, 190 Chromic acid, 245 Chromium, 103 Chrysoberyl, 74 Cinnamic acid, 68 Citrate, 229 Cobalt, 97, 103 blue glass, 77 glass, 53 spark, 142 Cocaine, 63 hydrochloride, 237 Cocoa, 193 Colemanite, 74 Collodion, 52 Colloidal metals, 50, 230 Copper, 88, 95, 97, 103 spark, 140, 142 sulphate, 54 Cordierite, 76 Cornea, 78, 206 Cotton, 245, 246 Coumarin, 228, 240 Crookes glass, 88, 119 Cultures, 237 Cyanine, 100 Cyanogen, 225, 231 bands, 160 Daylight intensity of, 17 photochemical action of, 33 INDEX 253 Detection of ultraviolet, 165 Diamond, 74 Didymium, 87 Digestion, 213 Di-saccharides, 232 Discharge tube, 134 Dyes, 52, 57, 67, 123, 243, yellow, 53 -testing, 242 Eclipse blindness, 205 Egg albumen, 208, 229 Egg globulin, 208 Electric waves, 5 Electromagnetic theory, 10 Emerald, 76 Enzymes, 216, 223 Eosins, 57, 67 Erythemia, 240 Eserine salicylate, 237 Esculine, 54 Esters, 59, 66, 227, 245 Ethane, 233 Ether, 52 Ethyl acetate, 62 Ethylene, 52, 225, 233 Euphos glass, 81 Exploding wires, 160 Eye, 78, 205 lens, 78 media, 206 Fatty acids, 59, 245 Fermentation, 239 Ferrous chloride, 245 Filters, 87, 184, 186, for mercury lines, 123 Flame arcs, in, 124, 242 Flames, 134 Flesh, 214 Flour, 228 Flowers, colors of, 220 Fluorenone, 61 Fluorescein, 58, 67 Fluorescence, 186 color of, 192 effect of solvent, 193 of various substances, 191 Fluorescent eyepiece, 190 solutions, 54 Fluorine, 234 Fluorite, 75 Fluorometer, 188 Fluorspar, 76 Formaldehyde, 226, 236 Formamide, 31 Formic acid, 229, 246 Fraunhofer lines, 23 Fresnel's formula, 81 Gamma rays, 5 Gases compound, 159 transparency of, 35 Gas-mantle, 109 Gasoline, 241 Geissler tube, 136 Gelatine, n, 77, 183 bichromated, 185 Gelatinization, 245 Gems, 75 Germicidal action, 209 Germs, 209 Glass blue uvial, 87 borosilicate, 82 chromium, 88 cobalt, 88 cobalt blue, 87 colored, 87 coloring media, 89 copper, 88 decolorizing, 86 effect of cathode rays, 86 effect of temperature, 85 effect of ultraviolet, 86 effect of X-Rays, 86 gold pink, 87 iron, 88 lead, 82, 88 lead oxide in, 83 manganese in, 85 nickel, 88 potassium oxide in, 83 reflection from, 81 254 INDEX reflection of, 99 soda, 82 sodium oxide in, 83 smoke, 89 transparent to ultraviolet, 83 uranium, 88, 89 uviol, 83 Glasses, 79, 119 colored, 84 crown, 79 eye-protective, 89 flint, 79 transparency of, 84 Glycerine, 51, 75, 238 Glycerol, 67, 238 Glycine, 65 Gold, 95, 99, 103 film, 119 spark, 140, 142 Grating, 153, 174 Grating spectroscope, 172 Grotthus* Law, 62 Guaiacol, 237 Gypsum, 74 Haematite, 137 Haloacetic acids, 241 Halogen, 67, 234 hydride, 235 Helium, 38, 136, 158 Hydriodic acid, 236 Hydrobromic acid, 229 Hydrocarbons, 42, 55, 227, 229, 234, 236, 245 Hydrochloric acid, 229, 236 Hydrogen, 37, 135, 227 arsenide, 233 chloride, 227 dispersion, 198 iodide, 227 peroxide, 66, 227, 228, 229 phosphide, 232 sulphide, 227 Hydrosulphite, 235 Illuminants, common, 107 Infra-red, 5 Inks, 244 Insects, 218 Iodide, starch, 234 Iodides, 235 Iodine, 99, 234, 235 derivatives, 236 lonization, volume, 40, 196 Ions, 224 Iron, 95, 115, 137, 153 arc, 161 arc in hydrogen, 147 arc simple, 146 oxalate, 229 spark, 141 Isoamyl, 233 Isopentane, 233 Ketones, 61, 66, 68, 227, 232 Ketoses, 226 Knowledge, early, 8 Kunzite, 74, 192 Kyanite, 76 Lead, 103 oxide, 83 phthalate, 239 Light, velocity of, 3 Limelight, 109 Linseed oil, 244 Liquids, transparency of, 46 Lithopone, 133, 243 Living matter, 204 Luminescence, colors of, 191 Magenta, 57 Magnalium, 95, 103 Magnesium, 103 burning, no iodide, 235 spark, 143 sulphide, 193 Magnetite arc, in, 115, 119 Malonic acid, 236 Manganese, 88 Measurement of ultraviolet, 165 Mercuric chloride, 234, 241, 245 Mercury, 99 arc, 53, in, 116, 119, 128, 130, 148 INDEX 255 energy in, 151 jacket, 229 benzoate, 237 bichloride, 237 lines, 121, 143, 152 filters for, 123 vapor, 43, 136 Metals, 94 disintegration of, 230 Methane, 231 Methylene chloride, 241 Mica, 77, 119 Microphotometer, 181 Micro-radiometer, 170 Milk, 215 Minerals, 75 Molybdenum, 104, 137 Monamines, 61 Monoses, 227 Moon, 17 Napthalene, 66 Natural gas, 233 Noedymium, 54, 87, 122 Neon, 158 Nernst glower, 109 Nernst lamp, 116 Newton's experiment, i Nickel, 88, 95, 99, 104, 139, 153 spark, 142 Nitrates, 67 Nitric acid, 67 Nitric oxide, 40 Nitrites, 225 Nitro compounds, 65 Nitrogen, 136 Nitrosodimethylaniline, 52, 67 Nitrous oxide, 30, 39, 225 Nuclei, 224 Oils, 67, 244 bleaching, 242 ethereal, 66 Olive oil, 237, 242 Ophthalmia, 205 Organic compounds, 227 Oxalic acid, 66, 185, 235, 236 Oxidation, 244 Oxidizing agents, 246 Oxycellulose, 245 Oxygen, 29, 30, 36, 135, 224, 228 Ozone, 29, 37, 39, 151, 224, 231, 236, 239 Paints, 244 Palladium, 104 spark, 143 Paraphenylenediamine, 127 Pathological effects, 117 Pentane lamp, in Peptides, 64 Peptone, 67, 229 Perchlorate, 231 Petroleum, 241 Phenol, 64, 65, 231 Phosgene, 242 Phosphorescence, 186 decay of, 188 Phosphorus, 152, 226, 231 Photo-electric cell, 101, 171, 194 spectral sensibility, 196 Photo-electric effect, 194 Photo-electric method, 47 Photo-chemistry, 223 Photographic density, 181 efficiencies, no emulsions, 183, 199 laws, 179 process, 180 value of radiants, 130 Photography by ultraviolet, 186 Photo-phthalmia, 205 Pitchblende, 137 Plant life, 218 Plants, 219, 240 Platinum, 95, 99 spark, 143 Polarimetry, 178 Polymerization, 225 Potassium, 99 bichromate, 123 carbonate, 193 iodide, 235, 241 metabisulphite, 66 256 INDEX nitrate, 241 oxide, 83 percarbonate, 229 permanganate, 123 sodium sulphite, 66 P-phenylenediamine nitrate, 185 Precipitating metals, 229 Proteins, 238 Purines, 63 Purplish tint in glass, 86 Putrefaction, 217 Pyridine, 65, 244 Quartz, 10, 35, 72 expansion of, 150 fused, 73 spectrograph, 175 Quinine, 245 bichloride, 237 sulphate, 123, 192 Quinol, 62 Radiation, 2 visible, 5 Radioactive substances, 67 Radiometer, 167 Radiomicrometer, 169 Reflection of ultraviolet, 93 Refractive index, 177 Resins, 244 Resonance spectra, 236 Resorcinol, 62 Retene, 52 Rock-salt, 75 Rontgen rays, 5, 10 Ross alloy, 95 Rubber, 237 Ruby, 76 Salicylaldehyde, 65 Saline solutions, 67 Saltpetre, 76 Salts, solutions of, 47 Sanidin, 74 Santonin, 63 Scandium, 157 Schroder's alloy, 95 Screens, non-selective, 180 Search-lights, 147 Sector-photometer, 181 Selenide, 227 Selenium, 99, 104 cell, 170 Serum albumen, 208 globulin, 208 Signalling by ultraviolet, 187 Silicon, 104 Silver, 95, 97, 153 bromide, no, 234 chloride, 74, 234 compounds, 9 film, 77, 123 fluoride, 234 nitrate, 237 spark, 140, 142 Silvered mirror, 68 Skin diseases, 217 effect on, 210 Skylight, 1 08 spectrum, 19 Sky, overcast, 108 Snow, 93 blindness, 204 Sodium, 99 cacodylate, 237 metabisuphite, 66 oxide, 83 sulphite, 66, 235 tungstate, 137 vapor arc, 161 Solar radiation, 15, 130 Solids, transparency of, 72 Sources, experimental, 133 Spark, 134, 136, 138, 153 apparatus, 136 gap, effect of ultraviolet on, 194 oscillating, 150 spectra, 155 spectral lines, 144 Spectrograph, 175 Spectrophotographic filter, 182 Spectroscope, 176 Spectrum, 2 INDEX 257 entire, 5 regions of, 5 Speculum, 104 Spodumene, 192 Spores, 210 Spinel, 76 Steel, 97, 100, 105 Stellite, 105 Sterilization, 212 Stilbene, 68 Stokes law, 193 Strontium iodide, 235 spark, 140 Styrene, 68 Succinate, 229 Succine acid, 236 Sucrose, 237 Sugar, 74, 227, 229 Sulphur, 232, 235 dioxide, 67, 229, 232 Sulphur fluoride, 234 trioxide, 237 Sulphuric acid, 66, 237 Sulphurous acid, 67 Sunlight, 1 08 duration of, 18 spectrum, 19, 23, 24 Sunstroke, 213 Sylvite, 76 Tantalum, 105 Tartaric acid, 236 Tartrate, 229 Tartrazine, 67 Telluride, 227 Tellurium, 105 Temperature, high, 161 Thermometer, 169 Thermopile, 167 Tin, 105 spark, 140, 142 Titanium, 115 Titanous chloride, 137 Toluene, halogen, 64 Toluol, 233, 241, 245 Topaz, 74 Tourmaline, 76 Toxin, 217 Transparency of of glasses, 79 of solids, 72 Tungsten, 105, 137 arc, 147 filament, 129 temperature of, 131 lamp, in, 114, 130, 182 -mercury arc, 130 Turpentine, 52 Ultraviolet radiation from various sources, 116 reflection of, 93 regions of, 5 Uranium glass, 88 Uranyl acetate, 185 Urea acid, 246 Vanillan, 240 Vapors of metals, 43 Varnishes, 244 Velocity of radiation, 3 Vinyl ester, 241, 242 Vitellin, 208 Water, 31, 47* JI 9, 215 absorption by, 49 color of, 50 -cooling, 240 transparency of, 217 vapor, 30, 39, 227 Wavelength influence of, 226 symbols of, 3 Wedge, 181 White pigments, 94 Wine, 238 Wires, exploded, 160 Wolframite, 137 Wood's alloy, 98 Writing, 198 X-rays, 5, n, 153 258 INDEX Xylene, 52 Yellow dyes, 67, 68 Zinc, 99, 105, 153 arc, 140, 146 ethyl, 231, 233 spark, 140, 141 vapor, 43 Zircon, 76 Zirconium nitrate, 158 OTHER BOOKS BY M. LUCKIESH COLOR AND ITS APPLICATIONS Second Edition, Revised and Enlarged. 6x9, 150 illustrations, 4 color plates, 431 pages . $4.50 The object of this treatise is not only to discuss the many applications of color, but to establish a sound scientific basis for these applications. The book is authoritative, well illustrated, and contains many references and a wealth of new material. It was written by an investigator in the general field of color and is therefore not narrowly limited in scope. It fills a distinct gap that has existed on the book shelves. LIGHT AND SHADE AND THEIR APPLICA- TIONS 6x9, 135 illustrations, 277 pages $3.00 The book is a condensed record of several years' research by the author in the science of light and shade. It is the first published work which deals with the science of light and shade hi a complete and analytical manner. The author has the faculty of bringing forth scientific facts in such a manner as to be helpful to those interested in the various arts. The book is of extremely wide interest because it deals with the appearances of objects and hence with vision and with lighting. It is well illustrated and represents the first elaborate attempt to formulate the science of light and shade and to correlate it with various arts. VISUAL ILLUSIONS, THEIR CAUSES, CHARACTERISTICS AND APPLICATIONS 6x9, 100 illustrations, 258 pages $3.00 There are numberless visual illusions, all of them inter- esting but many can be put to useful service in daily life. In this book will be found a condensed treatment of the practical aspects of visual illusions. The complexity of the subject is not overlooked but simplicity is attained by con- fining the treatment mainly to static illusions, by sup- pressing mere details and by subordinating theory. The book emphasizes experimental facts and introduces theo- retical considerations occasionally but chiefly for illustrating explanations which otherwise would be too complex. LIGHTING THE HOME 5 x 7f , illustrated, 289 pages $2.00 This is a pioneer book. It ranks with books on interior decoration and furniture as a help toward transforming a house into a home. It is practical in that it offers advice on all sorts of lighting problems and it is fascinating reading as well. ARTIFICIAL LIGHT, ITS INFLUENCE ON CIVILIZATION 6x9, illustrated, 366 pages $3.00 This story of the achievements of artificial light is written especially for the man in the street who is not interested in technical scientific terms and formulae, but who looks with admiration upon the huge signs which flash and sparkle above the crowds on the Great White Way, who marvels at the colors and brilliance of a spectacular theatri- cal production and desires to know how it is accomplished, and who takes a natural delight in hearing about scientific discoveries when they are explained hi the simple, vivid language he understands best. THE LIGHTING ART, ITS PRACTICE AND POSSIBILITIES 6x9, illustrated, 229 pages . $2.50 This book discusses lighting as engineering plus art, and treats the subject as a branch of interior and exterior decoration. The technical aspect of the subject is not neglected, but the main emphasis is upon the " why " and not merely the "how" of lighting. THE LANGUAGE OF COLOR 6x9, illustrated, 282 pages $2.00 A practical volume on color, the various fields in which it is used and its importance in portraying the ideas that make for progress. A book of special interest to all those who deal in color schemes and values. THE BOOK OF THE SKY 6x9, illustrated, 236 pages $3.50 " The beauties, wonders, awesome spectacles, inspiring panoramas, and extensive ranges of vision which await the aerial traveler, make of cloudland a veritable fairyland if he will open his consciousness to them. Aircraft have brought this new world of experiences within easy reach of mankind and it is one of the aims of this volume to awaken those who fly, or would fly, to the variety of interest which air travel affords." The Literature of Engineering and the Industries 4J On our shelves is the most com- plete stock of engineering, scientific, technical and industrial books in the United States. A large number of these we publish and for an ever in- creasing number we are the sole agents. ij All inquiries are cheerfully and carefully answered and com- plete catalogs as well as special lists are sent free on request. D. Van Nostrand Company PUBLISHERS AND BOOKSELLERS 8 WARREN STREET NEW YORK DATE DUE SLIP fN UNIVERSITY OF CALIFORNIA MEDICAL SCHOOL LIBRARY THIS BOOK IS DUE ON THE ^AST DATE STAMPED BELOW 2m-ll,'29