Digitized by the Internet Archive in 2007 with funding from Microsoft Corporation http://www.archive.org/details/condensevaporOObarurich CONDENSATION OF VAPOR AS INDUCED BY NUCLEI AND IONS By CARL BARUS Hazard Professor of Physics, Brown University WASHINGTON, D. C: Published by the Carnegie Institution of Washington May, 1907 LIBRAJTST VNIVERSITY OF CALIFORNIA •nAVTQ CARNEGIE INSTITUTION OF WASHINGTON Publication No. 62 PREFACE. The chief purpose of the present volume is the development of a fog chamber of simplest practical character, capacious enough to admit of the measurement of the largest available coronas, and efficient up to the highest exhaustions applicable; i.e., those which do not uselessly overstep the optical limits of the experiment, where fog particles become so fine as to be virtually inactive in diffracting or scattering white light. This I think has been accomplished, and the results, as far as they go, seem to indicate an efficiency not inferior to Wilson's piston apparatus. I have not, however, been able to get much beyond the large green- blue-purple corona, no matter whether the nuclei selected were effec- tively large, like the ions, or effectively small, like the colloidal nuclei. The forms beyond are flimsy and so nearly colorless as to be useless for measurement; but the steam-jet nevertheless reveals a whole order of axial oranges and yellows, lying beyond, which to my knowledge have not been detected in any form of fog chamber whatever. As used in most experiments, including my own earlier work, the fog chamber with a plug stopcock seems to be of very inferior efficiency in comparison with the piston form. This, however, in a properly de- signed apparatus, is the case only when the attempt is made to obtain the isothermal drop in pressure observationally at the fog chamber, closed at once after exhaustion. The datum needed can only be found by computation, and the initial pressures in the fog and vacuum chambers and their final pressure when in contact, always at the same temperature, suffice for this purpose. Though I was prepared for some corrections, I did not anticipate so large a difference between the apparent drop and the true drop of pressure, as actually appears. In the experiments which follow, the ratio is in fact as i,oco to 775, a difference of nearly 25 per cent. Hence it will be necessary to restandardize the coronas with this result in view, an undertaking which I hope to begin in the near future. Indeed a large number of incidental results would have made this desirable in the interest of other investigations. For similar reasons I have (as a rule) continued to refer the nucleations of the present volume to the drop in pressure observed at the fog chamber; and such reference is sufficient for the comparisons aimed at, if the same type of apparatus is used throughout, as was the case. Having improved the fog chamber to the degree shown in Chapters I and II, it was made use of in Chapter III for certain incidental experi- iii IV PREFACE. ments. I have already shown in case of dust-free air and the persistent nuclei produced by intense X-radiation that the distribution of nuclei within the fog chamber is a most remarkable feature of the experi- ment. The same, however, is true of the ions. Whether produced by X-rays acting from long distances or by radium, the density of ioniza- tion is as a rule very different in different parts of the fog chamber, showing the important effects due to the presence of secondary radiation within. Again, the change of nucleation produced when the exciting cause (X-ray bulb or radium tube) is removed at different distances from the fog chamber was to be reinvestigated. Associated with this experiment is the occurrence of minima of nucleation for certain dis- tances, supposing the exhaustion to be sufficiently high to induce con- densation of ions and colloidal nuclei in presence of each other. Finally, if the rate of decay of ions can be inferred from independent electrical experiments, a method for the standardization of coronas is presented which bids fair to be the most satisfactory solution of the problem suggested. The method admits of a determination not only of the rela- tion of the nuclei corresponding, in a given case, to two different coronas (cat. par.), but of the absolute nucleation involved. There is also a possibility of detecting in this way how a given mass of precipitated water is distributed among nuclei of different sizes when occurring together — one of the most important of the problems outstanding in connection with this apparatus. Chapter IV adduces a variety of results for colloidal nuclei in media other than air- water. It is shown, for instance, that there is no evidence to prove the colloidal nuclei in a medium of carbon dioxide and water are larger than in the normal case of air and water, in spite of the presence of the coercible gas in which groups of larger molecular aggre- gates would be anticipated. On the other hand relatively large colloidal nuclei do seem to occur in a medium of air and alcohol vapor. Thus it is suggested that colloidal nuclei in dust-free wet air are to be associated with the saturated vapor and that the gas is only secondarily involved. In Chapter V, undertaken by Miss L. B. Joslin under my direction, a systematic comparison is worked out of the relations between the num- ber of ions in the atmosphere and the corresponding dust contents in the lapse of time. No direct connection is apparent, whence it follows that as the nucleation is largely of local origin, other sources must be looked to for the ionization, or that the enormous local output of ions from an industrial community vanishes so rapidly as to be quite negli- gible. On the other hand, it may be possible to detect evidences of "absorption" in the curves obtained. Incidentally the nucleation of the atmosphere of Providence during nearly four years is exhibited. PREFACE. V Finally, in Chapter VI, I return to the problem begun in Chapter I, to see whether there is any change in the colloidal nucleation of dust- free air, such as might be ascribed to the ionization produced by some penetrating cosmical radiations coming from without; for it was to be the plan of these researches to study the ordinary dust content of the atmosphere with regard to its variation in the lapse of time first, and thereafter to continue in the same way with the nucleation of the dust- free atmosphere. This nucleation is found to increase synchronously with the decrement of the barometer; but as the amount of adiabatic cooling, i. e., the efficiency of the apparatus {ccet. par.), follows the same conditions, it is extremely difficult to disentangle the two effects. Never- theless the results are of considerable interest and they are therefore reported in their present state of progress. My thanks are due to Miss L. B. Joslin, who not only gave efficient assistance in the preparation of the manuscript and of the drawings for the press, but contributed much of the work in Chapter III, section 62, and Chapter V. Carl Barus. CONTENTS, Chapter I. — Early Successive Stages of Efficiency of the Fog Chamber and Allied Results. DISTRIBUTION OF NUCLEI WITHIN THE FOG CHAMBER ENERGIZED BY THE GAMMA-RAYS OF RADIUM, AND BY THE X-RAYS. Page, i . Introductory i 2. Data 3 3. Explanation 5 4. Further experiments with radium. High exhaustions 5 5. The same, continued. X-rays 6 6. The same, continued. Miscellaneous tests , 7 7. Distance effects of radium 9 8. Cause of the minimum and the maximum 10 9. Further experiments with radium 12 10. Distance effect of penetrating X-radiation 12 1 1 . General inferences (radiant fields) 13 THE NUCLEATION OF FILTERED AIR IN RELATION TO DIFFERENT SUPER- SATURATIONS OF WATER VAPOR. 1 2. Successive series of results 13 13. Effect of X-rays and gamma-rays. Data 18 14. Persistent nuclei 18 15. Persistent nuclei generated through tin plate 20 16. Discussion 21 17. More rapid exhaustion. Apparatus and data 23 18. Remarks on the j-curves. Non-energized air 25 19. The same, continued 26 20. The same, continued. Action of radium 27 II. The same, continued. Action of X-rays 27 22. Remarks on the n-curves. Non-energized air 27 23. The same. Action of radium 29 24. The same. Action of X-rays 29 25. Remarks on the iV-curves 29 THE NUCLEATION OF FILTERED AIR IN THE LAPSE OF TIME. 26. Method 30 27. Early data 30 28. Apparatus modified 31 29. Inferences 31 SUMMARY OF THE RESULTS OF THE CHAPTER. 30. Distribution of ions within the fog chamber 32 31 . Minimum of efficient nucleation '. 32 32. Persistent nuclei 32 33. Dependence of efficiency of fog chamber on the size of the exhaust pipes. ... 33 34. Invariable character of colloidal nucleation in the lapse of time 33 VII VIII CONTENTS. Chapter II.— Later and Final Stages in the Efficiency of the Fog Cham- ber due to a Gradual Increase in the Bore of the Exhaust Pipes. CONNECTING PIPES NOT LARGER THAN 1. 5 INCHES IN DIAMETER. Page. 35. Introductory 34 36. Examples of data for i-inch connecting pipes 35 37. Data for pipes 1 J inches in diameter 35 38. The same, continued. Shorter pipes 38 CONNECTING PIPES 2 INCHES IN DIAMETER. 39. Remarks on the method 4 1 40. Data for pipes 2 inches in diameter, 1 2 inches long 42 41. The same, continued. X-ray bulb inclosed in lead 44 42. Discussion 45 43. Radiant fields 49 COLLOIDAL NUCLEI IN DUST-FREE AIR. EXHAUSTION PIPES AND STOP- COCKS 4 INCHES IN DIAMETER. 44. Purpose 5 ! 45. Apparatus 5 1 46. Exhaustion difficulties 53 47. Same, continued. Case of air in fog chamber saturated with water vapor. . . 55 48. Case of saturated air in both chambers 59 49. Observations with 4-inch exhaust pipes 65 50. Observations, continued 70 5 1 . The same, continued 71 52. Discussion 73 53. Summary 73 Chapter III. — Miscellaneous Experiments. 54. Objects 77 55. Growth of persistent nuclei 77 56. Water nuclei produced by evaporation 78 57. Distance effects. X-rays 82 58. The same, continued. Small wood fog chamber 85 59. The same, continued. Large wood fog chamber 85 60. The same, continued. Discussion 86 61. Distance effect and absorption. Radium 89 62. Falling to pieces of ions in the lapse of time 91 63. Decay curve 95 64. The same, continued 97 65. Condensation phenomena of the inclosed steam jet. Methods and results .. . 98 66. Summary 102 Chapter IV. — Distribution of Colloidal Nuclei and of Ions in Media other than Air-water. COLLOIDAL NUCLEI AND IONS IN WET DUST-FREE CARBON DIOXIDE AND IN WET COAL-GAS. 67. Apparatus IQ 5 68. Data for carbon dioxide IQ 5 69. Behavior of carbon dioxide I0 7 70. Cause of differences IO j CONTENTS. IX Page. 71. Nucleation increases subject to a uniform law of equilibrium 109 72. Data for coal gas no 73. Character of the early results for coal gas no 74. New data for coal gas in 75. Conclusion 112 COLLOIDAL NUCLEI AND IONS IN DUST-FREE AIR SATURATED WITH ALCOHOL VAPOR. 76. Introductory 112 77. Apparatus and method 113 78. Properties of alcohol fog 113 79. Number of particles 113 80. Size of the nuclei 115 81. Data for alcohol vapor 1 18 ABSENCE OE COLLOIDAL NUCLEI IN STRONG ODORS. 82. Introductory 120 83. Data for camphor, turpentine, naphthalene 120 84. Summary 121 Chapter V. — The Cotemporaneous Variations 0} the Nucleation and the Ion- ization 0} the Atmosphere 0} Providence. By Lulu B. Joslin. 85. Introduction 123 86. Measurements of nucleation 124 87. Data for nucleation 125 88. Remarks on the table of nucleation 125 89. Mean daily nucleation 125 90. Mean monthly nucleations , 141 9 1 . Measurement of ionization 141 92. Data for ionization 143 93. Remarks on the tables 149 94. Errors of measurement 150 95. Mean daily ionization 152 96. Mean monthly ionizations and conclusion 153 Chapter VI. — The Variations of the Colloidal Nucleation of Dust-Free Air in the Lapse of Time. 97. Introductory 155 98. Method and data 155 99. Deductions 157 100. Effect of the barometer 158 101. Corrections 159 102. Further data 160 103. Conclusion 164 CHAPTER I EARLY SUCCESSIVE STAGES OF THE EFFICIENCY OF THE FOG CHAMBER AND ALLIED RESULTS. Before beginning the main subject of this chapter (see section 2), it is advisable to add a few measurements of the distributions of ions pro- duced within the fog chamber by the X-rays or by radium acting from without, since these occurrences must be kept in mind throughout the measurements. Again, the effect of different classes of nuclei (persistent nuclei, ions, and colloidal nuclei) in presence of each other is similarly important and direct light must be thrown upon it preliminarily. Under all conditions the fog chamber is attached to a large vacuum chamber by a rigid passage-way of the length and diameter specified, the ratio of the volumes of the two chambers being about as 6 to 100, respec- tively. Moreover, it was customary to read off the drop of pressure (dp) at the fog chamber (isolated immediately after exhaustion from the vacuum chamber) when isothermal conditions had been reestablished. The observed datum suffices for the comparisons of nucleation when the same chambers are used throughout; but it will be shown in Chapter II that it is much in excess of the true drop of pressure and that the latter is to be computed from the initial pressures in fog and vacuum chambers and the final pressure when both are in contact, all under isothermal conditions. After completing the work of section 2 of this chapter, a few appli- cations were made with the apparatus in its state of partial completion for the purpose of ascertaining whether there is any discernible change of colloidal nucleation and by implication of ionization in the stagnant air within the scope of the method. Several months of observation showed none. The early data have been added to the chapter for convenience in chronology, though they slightly interrupt the continuity of the research. The question of time variation is taken up again by a different method in Chapter VI. DISTRIBUTION OF NUCLEI WITHIN THE GLASS FOG CHAMBER, WHEN THE AIR IS ENERGIZED BY THE GAMMA-RAYS OR THE X-RAYS. 1. Introductory. — I may recall at the outset that there are three classes of nuclei to be considered in this chapter, the first of which includes the ordinary dust-like or persistent kind. They may be sepa- 1 2 VAPOR NUCLEI AND IONS. rated from the air by the filter and they require the smallest degree of supersaturation of water vapor to precipitate condensation. They are usually but not always (necessarily) foreign bodies in the air; at least they are producible in dust-free air by the X-rays of sufficient intensity and by other radiation. The second class comprises the fleeting nuclei. They are often charged and then called ions. They persist for very short periods of time, usually vanishing within a minute. They can be maintained, therefore, only in the presence of radiation, corpuscular or undulatory, from which their intimate association with electrification or with ultra-violet light is manifest. Such radiation may occur spon- taneously within the body of a gas during the state of generation. The sizes of these nuclei are intermediate between the first or dust-like class and the third class. This comprises the colloidal nuclei of dust-free air, which are virtually persistent, inasmuch as they are a structural part of the body of the gas and are reproduced as soon as removed. They require the highest degrees of supersaturation for condensation and are without electrification. All these groups may be made to pass continuously into each other. I shall use the term "nucleation" to denote the number of nuclei per cubic centimeter observed in any experiment. It will be understood that this means the efficient nucleation; for nuclei may be and usually are present, which are not detected by the exhaustion. They are com- puted in the present chapter from the angular diameter (**s/$o, nearly) of the coronas, for a given supersaturation. To specify their number one must come to a conclusion as to whether nuclei are removed more rapidly by the exhaustion than they can be replaced by the molecular system, or whether the reverse is the case. If fleeting nuclei and colloidal nuclei are supposed to be instantly reproduced we shall call the number n. None are then virtually removable by the sudden exhaustion. Otherwise the nucleation (corrected for the volume expansion) is called N. In this case the exhaustion is more rapid than the reproduction of nuclei. Both n and N, as well as s, will usually be given in the earlier tables; for the discrimination between n and N is not possible, and s is specially favor- able to the small nucleation which are apt to be crowded out of the diagrams for n and N. Returning to the subject of this section, I may state that the radium used was a weak sample (10 mg., 10,000 X) hermetically sealed in a thin aluminum tube. In an earlier paper I showed that whereas the nuclea- tion produced decreased very rapidly with the distance of the energizer from the outside of the cylindrical glass fog chamber, the number of nuclei within was apparently the same throughout the length of the axis. Pressure differences, however, were usually kept below the fog limit of DISTRIBUTION WITHIN FOG CHAMBER. 3 air. As the aluminum tube was carefully sealed (ground screw plug and wax) the beta- and gamma-rays are here alone in question, and as the bottom of the fog chamber through which the radiation passes may be 1 cm. thick and the walls are everywhere more than 0.2 cm. thick, only the more penetrating beta-rays are active to reenforce the gamma-rays. In fact the earlier work showed that the rays after passing 1 cm. of lead produced an amount of nucleation only 30 per cent less than in the absence of the dense barrier. Thus the whole phenomenon is practically a question of the intensity of the gamma-rays. In the course of the work, curiously enough a number of contradictory conclusions were reached, and it will therefore be advisable to report the results chronologically, beginning with the data for low pressure differ- ences. Table i. — Distribution of nucleation within the glass fog chamber (43 cm. long, 14 cm. in diameter); nucleator, radium in thin aluminum tube, dp = 21 cm. Radium at Dcm. from end, axially without. Lines of sight (two goniometers, j, and s 2 ) 15 cm. and 35 cm. from end nearest radium, or 20 cm. apart. D *i U A/, X 10- 3 io- 3 xA/ 2 Mean { $ Per cent. cm. 20 cm. 40 cm. co (air) 3-5 3-2 3-4 3-4 3-5 2-5 2.4 2.6 1.8 2-3 2-3 .0 3-3 3-o 3-4 3-6 3-3 2-5 2-5 2.8 2.0 2-3 2-3 .0 19 17 17 19 7 6 8 3 5 4 16 n 17 21 16 7 7 9 3 5 4 17,200 l6,200 7,000 8,000 4,000 4,000 100 45 24 2. Data. — The cylindrical glass fog chamber (fig. 1), rigorously free from leakage, was placed, with its axis horizontal, in such a way that a prolongation of the latter intersected the radium tube at a distance, D. Two goniometers with their lines of sight about 20 cm. apart, and 15 and 35 cm. from the end (bottom) of the fog chamber nearest the radium tube, were used nearly at the same time for the measurement of the apertures of the coronas seen along the axis. Both were placed with the pins nearly contiguous to the walls of the cylinder, so that the eye was at a minimum distance of 30 cm. off. All the phenomena may in this case be more clearly observed and the small coronas are less liable to drop out before the measurement has been completed. The source of light 4 VAPOR NUCLEI AND IONS. (Welsbach mantle) was 250 cm. beyond the fog chamber. Usually the angular diameter of the coronas is about 5/30. In table 1, s t refers to the goniometer nearer the radium, s 2 to the other, and A^ and N 2 are the corresponding nucleations, when D is the distance of the radium tube from the end of the cylinder. After the influx of filtered air, time was always allowed for the dissi- pation of convection currents. There was some difficulty in reading both goniometers consecutively without loss of time, as the small coronas soon vanish with the subsidence of the fog particles. It was necessary to cleanse the walls carefully before beginning the work to obviate deposits of dew. Fig. i. FIG, L -20 -10 Fig. 2. Fig. 3. ,2 1 c 1 *4 c £ ,1 c c 1 Fig. 4. Figs, i, 3, 4. — Types of cylindrical fog chamber. Axial section. E, exhaust pipe; c, cloth partitions. FiG- 2.— Nucleation (N) due to weak radium at different distances from and on top (side of cylinder) of the fog chamber. Table 1 . The table gives evidence of a small difference between N l and N 3 , but this is here unquestionably an observational error, particularly as its sign is often reversed. The chart (fig. 2) then shows the mean number of nuclei within, when the radium is at £> = o, 20, and 40 cm. from the end; and while the nucleation drops off rapidly from 100 to 45 and 24 DISTRIBUTION WITHIN FOG CHAMBER. 5 per cent, respectively, the nucleation within is about constant through- out the 45 cm. of length of cylinder. At least an internal drop to less than one-fourth is out of the question, as the diagram shows. It must be remembered that the nucleation is produced instantly to saturation, that all nucleation will vanish with the removal of the radium to infinity within a few seconds, and that the ions here in question are relatively large nuclei as compared with the colloidal nuclei of dust-free air. The latter, moreover, are quite ineffective at the observed pressure differ- ence, dp = 21, used. If larger pressure differences appear the results are usually quite different. 3. Explanation. — To account for these remarkable results is difficult. Convection is probably out of the question, though it will be quite elim- inated in the following experiments. One may hazard the suggestion that the effective agency outside of the cylinder are the gamma-rays directly, whereas within the cylinder the ionization produced by those rays secondarily is in question. In other words, the gamma-rays do not act here directly, but the nuclei are produced by corpuscles set free by these rays. The medium within the cylinder is thus in a state resembling an electrolytic medium having the same ionic pressure throughout. It is this ionic pressure depending on the density and speed of the cor- puscles which is transmitted instantaneously from end to end of the cylinder. To refer to the phenomenon as diffusion would be obscure without a statement as to what diffuses. The ionic nucleus is relatively a fixture and could not diffuse, in the time specified, to the far end of the cylinder, quite apart from decay. 4. Further experiments with radium. High exhaustions. — In the preceding experiments the wet cloth partition was about 10 cm. above the surface of the water below. In the new chamber shown in fig. 3, the distance has been reduced to 5 cm. Under these circumstances a very marked gradation of the number of efficient nuclei was observed for the first time. This is not, however, due to the supposed elimination of convection, but rather to the large pressure difference applied in the experiments. The result follows in table 2. When the radium tube is placed at T (fig. 1) symmetrically on the side, the metal cap virtually becomes the source of nuclei, or better, it seems to become secondarily active more intensely than the remainder of the glass chamber. The coronas obtained with two goniometers at distances 11 and 36 cm. from the brass end and 35 and 10 cm. from the glass end are rapidly larger as the brass end is approached. The ratios of the n- values is greater than 2:5. These conditions are retained indefinitely so long as the radium is 6 VAPOR NUCLEI AND IONS. present, showing that the result is not incidental. As both ends of the chamber were about equidistant from the radium, it would appear to be the excess of the secondary radiation from the metal cap which is in question. Table 2. — Distribution of nucleation within the glass fog chamber (as in table 1) and allied results. N xio -3 . N Xio- 3 . I. Radium tube on side of glass fog chamber. Two goniometers, I at 11 cm. and II at 36 cm. from brass cap, 46 cm. from glass end. dp m 26.0 cm. '{ No. I No. II No. I No. II 6.6 95 4-7 38 6.6 95 4.4 30 [47 59 '47 46 II. Radium on glass end, 1 cm. thick; observation I at 10 cm. and II at 35 cm. from end. dp = 26. No. I . No. II. 60 47 76 38 59 A/Xio- 3 . AT xio- 8 . III. Radium at 00 . Lapse 10 m . 26.5 30.8 2 7-4 11 114 17.4 195 IV. Radium on brass plate, 1 cm. thick. 30.6 4.0 24 42 1 Three hours later no interference. 2 wrg. 5. The same, continued. X=rays. — After finding that the fleeting nucleation within the fog chamber was unequally distributed under the excitation of radium, similar experiments were tried with the X-rays, using the fog chamber (fig. 4) with four horizontal wet cloth partitions 2 to 3 cm. apart to obviate convection. The results given in the first part of the following table are definitely affirmative. When the X-ray bulb is near the fog chamber, the nucleation nearer the bulb is decidedly in excess. Thus when the line of sight is 30 and 50 cm. from the anti- cathode, the farther nucleation is about 60 per cent of the nearer. If, however, the bulb is relatively far away (D = ioo or 200 cm.), the end of the fog chamber farthest from the bulb and near the brass cap is more actively ionized. In other words, under these circumstances the secondary radiation from the brass cap predominates. The nucleation at 10 and 30 cm. from the latter shows a decrement of 20 per cent, for DISTRIBUTION WITHIN FOG CHAMBER. 7 instance, passing continuously from the high to the lower values. This experiment throws light on the function of the walls in producing dis- tributions in case of persistent nuclei. Obviously the effect of secondary radiation decreases rapidly as the distance from the walls increases. The nucleation within the fog chamber is probably largely due to this kind of radiation. 6. The same, continued. Miscellaneous tests. — In the second part of table 3, experiments with radium are resumed for comparison. The gradation of number, decreasing from the brass end to the glass end, for a symmetrical position of the radium tube at T, fig. i, is very marked, as usual; while the case for radium at R, fig. i, again shows the largest coronas at the end of the tube, the distribution being more nearly uni- TablE 3. — Distribution of ions within glass fog chamber. Metallic (brass) cap at one end, glass end opposite 1 cm. thick. Lines of sight 10 cm. (s 10 ) and 30 cm. (s x ) from brass end, 15 cm. and 35 cm. from glass end. X-rays acting from distance D. Lead case. D. "^10- **>■ N i0 X 10- 3 . N„ X 10- 3 . dp. cm. I off 15 15 GBP WPcor 130 180 156 255 24-5 30 w P cor wrg 15 w P cor wog 180 300 30 1 w P cor wog 180 300 30 hoo 6.9 w 6.3 wg 121 9i 30 II off: ^00 6-5 6.1 102 84 30 On middle of side at T. •••{ 6.8 6.6 5-4 5-o 114 106 63 48 30 30 On glass end at R. . . On side near cap at S. On middle of side at T. On side near glass end 5-6 4-4 6.4 5-2 3-3 4-5 69 33 96 56 H 36 30 30 30 at S 1 4.8 4.8 45 45 30 Large coronas near brass cap. Secondary radiation preponderating. form. Thereafter the radium was placed in positions 5, T, S\ on top, successively, with the results for T graded about as usual. Radium at 5 still evokes gradation, but the coronas are even much smaller than when the radium is placed on the thick glass end at R. When radium is placed at S' near the glass end gradation is nearly absent, but the coronas are again small. The results are therefore not as simple as was antici- pated and do not admit of an explanation merely in terms of the glass penetrated. 8 VAPOR NUCLEI AND IONS. When the radium tube is attached to the glass end (about i cm. thick), similar gradations are observed, but this time with the large coronas near the radium tube. (See fig. 5.) The ratio of n- values is now about 2 for the given distances. The law of variation is difficult to obtain in any of these cases, but when the brass end is predominatingly active, the reduction of n- values is nearly proportional to distance, if the usual mean values of 5 be taken together with the special observa- tions. 120 80 60 20 \ m Rad. brass end DISTRIBUTION WITHIN FOG VA V*. ***j* V Radid/n on $i)sfe 6p >D &N FIG 26 CHAMBER w / /- **£ <**•" J* ^^•O - Radiurr ong/as. I end, op =26 <• ^ o ^ A0$ I CI 20 Fig. 5. — Density of ionization (A7) within the fog chamber at different distances (D) from the brass end. Table 2. In view of the action of the glass end when the radium is attached to it, it seems possible that the glass, large thicknesses of which are always penetrated and permeated obliquely by the radiation, may be active in both cases, but special inquiry is needed here. The glass cylinder was found to be thinner (about 0.15 cm.) near its equatorial parts (7), and to increase in thickness towards both ends. If, how- ever, the phenomena were due to thickness of glass the coronas should be largest near the middle and taper down toward both ends of the fog chamber. This is not the case, as the coronas for a symmetrical position of the radium tube decrease regularly from the brass to the glass end of the apparatus. If the radium tube is placed on the outside of the brass cap, one cm. thick or more, there is the usual moderate reduction of nucleation here, from n = 6$ to ^ = 24 thousands (63 per cent); but the change of diameter of coronas within is of the same kind, decreasing from the brass to the glass end, as above. DISTANCE EFFECT. 9 7. Distance effects of radium. — To throw further light on present and preceding questions at issue, I shall next anticipate some of the work below and describe certain experiments specially adapted for the pur- pose. By "dust-free air" I mean atmospheric air filtered with extreme slowness (through large wide filters of packed cotton) and thereafter left without interference for two or more hours. Such air shows a high fog limit. In the fog chamber used the condensation began at an observed pressure difference of about dp = 26 cm.; rain-like condensation at dp = 21 cm. In the present experiments all tests are made at dp = 41.$ cm., at a pressure difference therefore much above the fog limit, and probably approaching the condensing power of the apparatus. The number of nuclei computed from the coronas observed is an estimate merely, as the constants needed for the very large range of variation in question are not available. Nevertheless if the same dp is used throughout, the nucleations obtained are immediately comparable. With these reser- vations the number of nuclei found in the dust-free air and at the dp in question is about N = 380 X io 3 to 400 X io 3 , per cubic centi- meter. It is obvious, moreover, that these nuclei are excessively small, much smaller than ions, smaller even than those which would respond to smaller dp's exceeding dp = 26 cm. Let the fog chamber (fig. 1) be subjected to the radiations from weak radium (10,000 X, 10 mg.) contained in a thin hermetically sealed aluminum tube. As the walls of the fog chamber are 0.1 to 0.3 cm. thick and the end (bottom) toward the tube nearly 1 cm. thick, y-rays only will penetrate into the inside, apart from the secondary radiation there produced. As shown in fig. 1 (cylindrical fog chamber) the radium tube R is at an axial distance D from the nearer end. In addition to this the radium was also tested at T (top) in the figure, where it is nearest the body of dust-free air under experiment. The data investigated are shown in table 5 and in the curve (fig. 6), where the abscissas are the distances D and the ordinates the number of efficient nuclei per cubic centimeter. It follows from the graph that as the radium is brought in an axial direction from 00 to the end of the fog chamber, the number of efficient nuclei in the dust-free air contained is gradually but enormously reduced to a minimum for D = 2$ cm. (about), after which the number again increases to the maximum at D = 0. Curiously enough, when the radium was further approached to the body of the air by being placed at T, the mean number of nuclei did not increase.* *This result is to be further interpreted in special experiments below. IO VAPOR NUCLEI AND IONS. If the radium is inclosed in a long, thick lead tube (60 cm. long, walls 0.5 cm. thick), the nucleation is but moderately reduced (see crosses in fig. 6), showing that gamma-rays are in question. Table 4. — Radium effect at high pressure differences, dp =41.5 cm.; s (air) =7.5; D, distance from end of fog chamber. D. s. n X10- 3 . D. s. n x 10- 3 . On top. . 3-9 27 50 4-8 51 3 9 27 4 •9 54 25 3 3 16 100 6 3 104 3 4 17 6 8 129 4 8 51 00 7 ,1 160 4 7 48 End 4 8 51 On top. . 4 27 4 7 48 4 1 31 Top 4 6 44 4 4 37 4 7 48 4 7 48 End 4 7 48 25 3 3 16 Top 4 7 48 >o 4 1 3i 2 End 4 7 48 On top . . 4 27 00 6 97 x On top.. 3-9 27 1 Radium tube in lead pipe walls 5 cm. thick. * Lead radiation. 0& ftp** FIG. 6 J, 'AN) (JO -3 y\ h V -cfc^^ij %*JQ~* fp-22 *»D 40 60 do 100 120 Fig. 6.— Efficient nucleation (JV) near the middle of the fog chamber, for different dis- tances (D) of radium from the end and for different exhaustions (dp). Table 4. 8. Cause of the minimum and the maximum.— This is easily explained, since the ions are relatively large bodies and relatively few in number as compared with the nuclei of dust-free air for the same dp. Hence the ions capture the moisture more and more fully as their number, with diminishing distance D, becomes greater. At # = 25 cm. probably the whole of the moisture is condensed on ions, and as their number increases as D vanishes the minimum in question results. In fact, it DISTANCE EFFECT. II was shown elsewhere that below the fog limit of air, the nucleation observed and due purely to radium at different distances, D, is, for example (dp = 2 2), D N x io~ 8 o 20 10 50 30 32 50 100 20 12 etc., agreeing in character as far as may be expected with the data here in question. These data multiplied by 4 (in other words 4n X io~ s ) are also given in fig. 6 for comparison. Hence the ions caught at dp = 41.5 are about four times more numerous than at dp = 22, and corre- spondingly smaller. They behave, therefore, as if they were markedly graded, but nevertheless, as a group, throughout much smaller than the nuclei of dust-free air so long as the radiant field is appreciable. While the number of nuclei continually grows smaller, with dimin- ishing D, the efficient nuclei may nevertheless increase again below a certain D, seeing that the colloidal nuclei in dust-free air are enor- mously in excess, only a few of which are caught even in the absence of radium. Table 5. — Distance effect of radium. D measured from side. Sp~ 35 cm. D. s. n x io~ s . n xio -3 . cm. 250 200 150 100 50 25 12 cm. »7.6 '7-6 > 7 .6 s 73 4 5-i 4.4 4.2 4-7 160 150 145 136 56 35 30 45 300 285 270 255 105 66 57 84 Radium at 00. Lapses 15s and 3 m . 00 00 4.1 l 7-6 27 160 5i 300 'GBP. 2 gyog'. »wog. 4 wrg. It would be more difficult to account for the result that the same nucleation is observed wherever the radium touches the outside of the elongated fog chamber. In otherwords, radium at the end of the chamber in the earlier experiments seemed to produce the same mean nucleation as when at the top, although the distances from the center of mass of the glass are as 3 to 1 ; but the above experiments have already sug- gested that this result was probably an incidental case. 12 VAPOR NUCLEI AND IONS. 320 240 9. Further experiments with radium. — Incidental experiments corre- sponding to those of the preceding section are given in table 6, showing the reduction of the efficient nucleation of dust-free air when the radium tube is gradually approached to the sides of the chamber. They are shown graphically in fig. 7, both in relation to s, ft, and N, and need little discussion. The minimum position lies very near the chamber and about 10 cm. from it. The action of radium is hardly perceptible beyond 1 meter, chiefly because the green-blue- purple corona happens to be involved. Sharper re- sults could therefore be obtained at lower pressure differences, where white- red-green coronas are in question. But on the whole it is clear that ra- diation which appreciably lowers the asymptote cor- responding to dust-free non-energized air must be of an intensity exceeding the effect of 10 mg. of weak radium (10,000 X), at a distance of 1 meter. There not, therefore, seem to be much hope of ob- serving cosmical radiations in this way. Finally the second part of table 6 (and other data below) shows how soon the high nuclea- tions of dust-free air (non-energized) is regained after the removal of the radium tube. The rapidity with which ionized nuclei fall apart is obtainable in this way and well worthy of special research. (Cf. Section 62 et seq.) 40 uxp- 3 does Fig. 7. — Efficient nucleation (N) and coronal aper- ture (S) for different distances (D) of radium from end of fog chamber and for different exhaustions (dp). Table 5. 10. Distance effect of penetrating X-radiation.— The question at issue was whether the rays which have penetrated lead are characterized by the same marked distance effect which is observed for radium. The fog chamber was inclosed in a close-fitting lead casket with two narrow side-windows for observation. The end of the casket toward the bulb DISTANCE EFFECT. 13 was provided with a lid to be opened or closed at pleasure. When the distance is D = $o, about 8 per cent of the radiation passes through; when D = io cm. the amount is about 17 per cent. Unfortunately the observation through the window is not satisfactory and no doubt a certain amount of radiation enters here secondarily. It is therefore difficult to carry out the comparison. Tabids 6. — Penetration through lead. Lead-cased glass fog chamber. Plates of lead 0.14 cm. thick. dp= 24.5. Lead case — D. s. n X io -s . On ; end closed 50 50 10 10 10 15 15 2.7 3-1 gBP. 3-8 3-5 3-5 wog 7.0 j gBP to \ w P cor. 6.7 9-5 too 19-5 16.5 16.5 104 130 156 On; end open On ; end closed On ; end closed On ; end closed Off ; plate only Off ; no plate Large coronas near brass cap. Secondary radiation. 11. General inferences. — The occurrence of a continuous succession of groups or gradations of nuclei in the curve of figs. 6 and 7, each of which groups constitutes a condition of chemical equilibrium for the given radiating environment, is suggestive. In the first place, it may be recalled that the nuclei of dust-free air are an essential part of this body as much as the molecules themselves. Such nuclei, if withdrawn by precipitation, are at once restored. Again, air left without interference for days shows a maximum of this nucleation for the given conditions of exhaustion when all foreign nucleation must have vanished. Indeed, the water molecules themselves may be treated as a continuous part of the nucleation in question, the frequency of occurrence being a maxi- mum for the molecular dimensions. It is best, however, to reserve dis- cussion for a later paragraph (section 43). THE NUCLEATION OF FILTERED AIR IN RELATION TO DIFFERENT SUPERSATURATIONS OF WATER VAPOR. 12. Successive series of results. — Before proceeding with an account of the change of the efficient colloidal nucleation of atmospheric air in the lapse of time and other like problems, it is expedient to consider the behavior of filtered air when subjected to different exhaustions. I will I4 VAPOR NUCLEI AND IONS. proceed chronologically, taking the earliest experiments first. These may be given with sufficient detail in a chart like figure 8, in which the lower graph (I) shows the data obtained in the earlier memoirs with a wooden fog chamber. The observed pressure differences (super- saturations) are laid off horizontally, the apertures of the coronas (5 = ^/30, nearly, where is the angular radius to the outside of the first ring, when the eye and the source of light are at distances 85 and 250 cm. on opposite sides of the fog chamber) and the nucleations vertically. No particular attention was given to the occurrence of rain, and the measurements cease at the lower end with the measurably visible coronas. It is seen that the curve very soon reaches an asymptote, showing hopelessly poor efficiency. As the work was carefully done, the only explanation would seem to be that the speed of filtration was insufficiently slow and the exhaustion insufficiently rapid (small vacuum chamber). The same remarks apply to the graph (figure 8, II), ob- tained in a decreasing march of pressure difference, at the beginning of the work with the glass fog chamber (fig. 1); but the asymptote is higher. The fog limit in this case is lower and attributable to nuclei which have entered unobserved, as it is much below the values for more carefully filtered air. The first observation, having been made with the apparatus at rest for 12 or more hours, shows a very high but probably more nearly correct result. Figure 8 also contains the corresponding nucleations, N, so far as they can be computed, supposing that the nuclei are removed faster than they can be restored. For high pressure differences the data are necessarily estimates. Naturally the differences between the data obtained after long waiting (12 hours or more), and those obtained in succession after intervening intervals of a few minutes, are enormously accentuated. It follows from these results that if normal data were to be reached it would in the first place be necessary to perfect the method of filtration. This was done by providing the filter with a very fine screw stopcock, by which the rate of flow through the filter could be diminished very gradually to a vanishing value. The next group of results, obtained on May 21, 23, and 30, and on June 3, were found under conditions of very slow filtration. They are sufficiently reproduced in the graphs (figs. 9, 10, 1 1) ,in which the abscissas are the pressure differences and the ordinates the coronal diameters (s) and the nucleations (n and N) t respectively, as specified. The graphs for 5 are perhaps most suitable for discussion, although s exhibits the small nucleations with very great advantage. The data for May 21 are in the main much above those in fig. 8, but they are still far too low and EFFICIENCY OF FOG CHAMBER. 15 irregular. The other data are much larger and as far as dp = 30 lie sufficiently near together to show that limiting conditions are being approached. Beyond this the curves diverge widely. The highest results reached are those of May 30. On June 3 a change takes place during the course of the experiment, as the result of which the incoming and outgoing curves differ. In all cases there is a distinct tendency to approach a limit after dp exceeds 37 cm., due to the limitations of the apparatus. Figs. 8, 9, and 10. — Early data of efficient nucleations (N, n) and coronal diameter (j), appearing at different exhaustions (dp) in dust-free air. Imperfect fog chamber. The characteristic feature of all the nucleation curves is the rapid increase between dp = 27 and 30 cm.; that is, during the early coronal stages. Below this the relatively large nuclei are present at least as far as dp = 21 to the extent of a few thousand per cubic centimeter. Never- theless, it is their variation which brings about the sinuous outline of the curves above dp = 33, seeing that large nuclei are overshadowingly effective. i6 VAPOR NUCLEI AND IONS. On the 5th and 6th of August a series of experiments was made to test the effect of long and short lapses of time between the observations, as well as the displacements of lower limits of the curves. The results of August 6 at ^«=35 show that 15 minutes was an insufficient time between observations, but that intervals exceeding 30 minutes (very 600 500 NxlO' 9 Figs. Il, 12, 13, and 14. — Early data of efficient nucleations (A/, n) and coronal diam- eter (s), appearing at different exhaustions {dp) in dust-free air. Imperfect fog chamber. slow filtration presupposed) are liable to show the true efficient nucle- ation at the time of experiment. Direct experiments made under con- ditions show that intervals of from one to two hours will in all cases be sufficient for the decay of extraneous nuclei. EFFICIENCY OF FOG CHAMBER. 17 Table 7. — Nucleation of dust-free, non-energized and energized air, different exhaus- tions, dp. Lapse about 15 min. Barometer, 75.97 cm. Piping and stopcock of 1 -inch gas pipe. dp. s. n x 10- 3 . N X io- s . dp. s. n X io- s . N x io- s . I. 1 Dust-free air not energized. IV (continued). X-ray bulb. D = 100 cm.; fronting bottom. 4 35.1 26.0 6.2 2 94 27 180 3-9 22.7 5-6 59 85 26.3 24.2 2-3 1.2 3-8 1.4 5-8 2.1 20.9 18.8 4-9 •?. 9 ?I.O 1 . 2 37 •9 53 ?i.i ?1.2 1.6 22.6 22.6 21 .O (?) 1:2 (?) .2 .0 (?) •3 .0 19.0 19.0 •9 1 . 2 V. X-ray bulb. D = 50 cm. II. Ra< iium on middle of s chamber. ide of fog 19.0 1 .2 1 .2 1-7 VI. X-ray bulb. D = 25 cm. 21.0 19.0 19.0 3-3 (?) 1 .5 11 (?) .2 15 (?) .3 19.2 « 1.8 2.4 20. 1 2. 1 25 3-4 19.2 1-3 1.8 22.4 24.1 4.8 4-9 37 4i 53 61 18.2 18.2 .0 .0 .0 .0 .0 .0 26.0 5-2 52 80 30.0 5-2 55 93 VII. D=$o cm. above 8 fog chamber. 35-0 51 56 106 1.6 19.0 1 V 1 . 2 III. Dust-free air. 19.9 2 9 6.7 91 20.9 4 4 26 36 35o 6.2 94 180 21.3 5 6 54 77 33-0 5-8 80 *43 22.7 5 8 63 9i Sl.l 5-5 68 117 24.4 wrg 7 3 117 175 29- 3 5-2 54 9i 26. 2 wog^ 8 130 200 27.6 3 1 10 16 27.7 wo / g7 8 135 217 28.1 3-8 21 34 29.8 gBP8 3 150 249 26.3 2.1 3 4-7 9 33-0 gBP 156 281 28.1 3-8 21 34 34-9 gBP 160 303 29.6 4-7 4i 70 38.3 gBP 167 346 31 -2 5-3 5-8 60 80 103 42.0 wog> 7 . 9 ?i 7 4 ?4QO 33- 1 143 35-2 s 5-8 82 155 VIII. D=6oo cm. from end. IV. X- ray bulb. D = 100 cm .; fronting 35-0 10 4-4 35 67 bottom. 4 33-7 4.6 4.6 4i 75 68 31.0 40 42.0 5 7-o 146 338 28.0 4.6 37 60 41.4 '7.0 H5 332 25.8 4.6 36 56 35o 6 7-o 134 255 24.2 4.6 35 52 30.7 5 7-2 131 222 22.7 4-3 26 38 26.2 6-5 94 146 20.9 3-o 7-7 11 24.0 6.2 79 "3 19.7 i-3 i-3 1.8 1 Vanishing at once and faint, only just visible. * Nothing (fog or rain) observed. ' Accidental slow reopening : same result. ♦Exposure a minutes. No increment of N results. * All large wrg coronas. •Coronas clear but too small. 7 Bottom of glass i cm. thick, interferes with for- mation of persistent nuclei even at short distances like 25 cm. = i>. •Persistent nuclei just begin to be generated. a < 0.5 cm. maximum for fleeting nuclei. 9 Next day. 10 Smaller than air coronas ; a always taken. i8 VAPOR NUCLEI AND IONS. 13. Effect of X-rays and gamma=rays. Data. — To interpret the above results it will be necessary to vary the nucleations of the dust-free air artificially, by acting on it from the outside of the chamber with the X-rays, or with the gamma-rays of radium. In these experiments the glass fog chamber (fig. i), rigorously free from leakage, was used with a framework of wet cloth within. Filtration was throughout excessively slow and all other precautions were taken. The radium in the hermetically sealed aluminum tube was pasted to the top of the fog chamber at T. The X-ray bulb was adjustable, so as to act from different positions at distances D. Table 8 shows results where dp is the pressure difference, 5/30 the angular coronal diameter, when the distances of the eye and the lamp are 35 cm. and 300 cm. from the fog chamber. The small distances D (35 cm.) enabled the observer to see even the smallest coronas distinctly. The nucleation n is obtained by supposing that the nuclei are restored faster than they can be removed by exhaustion. In case of N this is not the case, so that the correction for volume expansion is added. For high pressure differences the absolute numbers are not trustworthy, but the data given nevertheless show the relations satisfactorily. Table 8. — Persistent nuclei. Exposure 2 m . D measured from anticathode to top of fog chamber. Thick (1 mm.) aluminum screen (earthed) interposed. D. dp. s. N x io -8 . Lp. D. dp. s. N x io~ 3 . Lp. cm. cm. min. cm. cm. min. 12 18.0 7.8 140 30 18.2 3-o 9 1 20 18 6 5-3 56 18.2 3-2 12 18 4 5-7 67 1 50 18.4 .0 30 18 2 40 18.3 2.7 7.2 1 18 *4-o 22 1 18.0 1 .0 1 . 2 1 17 5 3.o 9 2 18.0 1 .0 1.2 1 18 2 2.7 7 50 18.0 *a . 1 2 18 2 l 3-i 10 1 •5 . 1 1 1 Spontaneous generation. 8 Just visible before falling. The data of table 8 are constructed in the charts, figs. 1 5 to 1 7, showing the 5, w, and N curves in succession, for increasing supersaturations. They will be useful for comparison with the advanced work below. 14. Persistent nuclei.— Table 8 gives the corresponding data for persistent nuclei. Distances are here measured from the sides of the fog chamber where the glass was thinner (about 0.2 cm.) than at the end (1 cm.). No persistent nuclei were obtained when the radiation EFFICIENCY OF FOG CHAMBER. 19 passed through the end of the fog chamber as in table 7. To guard against direct inductive action the fog chamber was covered with an earthed plate of aluminum, 0.5 mm. thick, though this precaution is superfluous. The table shows the distance (D) of the anticathode above &BP Figs. 15 and 16. — Early data of efficient nucleations (N, n) and coronal diameter (s), appearing at different exhaustions (dp) in dust-free air. Imperfect fog chamber. (Illustrating table 7.) the fog chamber, the coronal diameter (5/30), the nucleation (N; n would have no meaning here) , and the time elapsing (Lp) , after the radiations are cut off until the observations are made. The time of exposure was uniformly 2 minutes. Longer exposures would have produced seriously 20 VAPOR NUCLEI AND IONS. distorted coronas. These data are constructed in fig. 18, the s values being dotted and the TV values drawn in full. 120 300 200 Fig. 17. — Nucleation (N) for different exhaustions (dp) and radiations. Bulb distance. Fig. 18. — Persistent nucleation (N) produced by X-rays acting from different distances (D) from the glass fog chamber. Table 9. 15. Persistent nuclei generated through tin plate. — In view of the ease with which the cylindrical fog chambers may be made rigorously air-tight, it seemed worth while to endeavor to produce persistent nuclei through a plate of tinned sheet-iron, 0.03 cm. thick and 14 by 20 inches square. In the following trials exposures for 2 minutes to the EFFICIENCY OF FOG CHAMBER. 21 X-ray bulb, at a distance of about 12 cm. between the anticathode and the glass wall (0.3 cm. thick) of the fog chamber, are in question. The earthed iron plate was placed between bulb and chamber, resting on the latter. A lapse of time of 1 minute was allowed after the radiation had been cut off. This suffices to show that nuclei of the persistent kind have been entrapped. Naturally the pressure difference used precip- itated no appreciable fog in the absence of radiation. These results show conclusively that persistent nuclei may be pro- duced through earthed tin plate, provided the thickness used is not excessive. Table 9. — Persistent nuclei through earthed tinned iron plate, 0.03 cm. dp*- 25 cm., practically below fog limit. Exposure to X-rays, 2 minutes; lapse, 1 minute; obser- vation during exposure. X-rays — s. n X 10- 3 . On Off On 2.0 .0 2.0 1 Weak 2-5 .0 25 .0 Off On { dp= 24.5; corona too faint for measurement, but definitely present. 16. Discussion. — The simplest of the graphs are those of figure 18, showing the decrease of persistent nuclei as the bulb is gradually removed (D increasing and measured from the side of the chamber) from the apparatus. This decrease is very rapid and implies that if the anti- cathode were actually in contact with the glass walls the production of persistent nuclei in 2 minutes would be enormous. The distortion of the coronas resulting under these circumstances, as detailed in the last memoir, is not, therefore, surprising. It should be remembered that the rays have to penetrate more than 2 mm. of glass. On the whole the action is very much like what C. T. R. Wilson describes for ultra- violet light. As in the earlier case, there is secondary generation; i.e., the number of nuclei is larger if the observations are made at 1 or 2 minutes or more after the radiation has been cut off. With the given apparatus, persistent nuclei were still produced for a distance of 0.5 meter between the X-ray bulb and fog chamber, though the coronas in that case were but just discernible and vanished too rapidly for measurement. The corresponding intensity of ionization 22 VAPOR NUCLEI AND IONS. is then near the point of transition from persistent to fleeting nuclei, a result which will presently be made use of. The pressure difference (dp = 18 cm.) used throughout is (for the given apparatus) far below the fog limit for dust-free air (dp = 21 to 26). In case of the fleeting nuclei or ions, the generation is instantaneous and it will be necessary to represent the data (s, n, N) for a continuous series of supersaturations for each of the five intensities of ionization applied, beginning with non-energized dust-free air. These are given for both increasing and decreasing values of dp, the latter being at times larger (as would be expected from the removal of larger groups on con- densation) ; but on the whole the data are very satisfactory. Remembering that the number is roughly as the cube root of s, the region of ions (dp = 21 to 26) is particularly well represented. They increase slowly but regularly in number until the coronal stages are reached. Beyond these the increase is again slow; but that it is real is shown by the top curve for X-ray nuclei, which fails to ascend. The lowest pressure at which the suggestion of a corona could be observed in the given apparatus was dp = 21 cm. The coronal fog limit is best found from the other curves. The succeeding curves (X-ray, D = 600; radium, D=o; X-ray, .D = ioo, 50, 25 from end and 50 cm. from side) correspond to gradually increasing ionization. It may be noticed that the fog limits slowly move to the left into smaller supersaturations, while the initial slope of the curves gradually rises. Persistence sets in when the slope is nearest the vertical. All the curves eventually approach a limit, which for weak radiation is practically reached near the fog limit of air, but for strong radiation much beyond it. In the upper curves the green-blue-purple corona first appears at dp = 30 cm. It is noteworthy that the asymptote is reached later as the ionization is stronger. The relation of these results as to numbers of nuclei is best shown by the w-curves, fig. 16, though the scale is now too small for the ions of non-energized air. In this case (n) it is assumed that the nuclei are reproduced faster than they can be removed by expansion. The curves throughout have the same characteristic, being doubly inflected and showing definite stages in which nucleation increases most rapidly. If these branches are prolonged backwards the coronal fog limit may be specified for each, as shown in the figures. This quantity decreases at a retarded rate, while the ionization grows more intense. The effect of weak ionization (radium, or X-ray at D = 6oo) is to increase the number of ions. It has been supposed above that the result of this is to mask the corresponding increase of any of the col- loidal nuclei originally present, should such an increase occur. EFFICIENCY OF FOG CHAMBER. 23 The curves, moreover, throw important light on the distance effect D. Thus at dp = 22, the effect of increasing D from 1 to 6 meters is a fall of n from 50,000 to 20,000 — about 60 per cent. At dp = 30 the fall is from 130,000 to 40,000 — about 69 per cent. Owing to the uncertain- ties of observation these numbers may possibly be the same, in which case the gradation of size of the nuclei would be the same for all inten- sities of ionization. The initially slow increase of the upper curves (D = 5o cm.) is possibly due to the formation of persistent nuclei, as stated at the beginning of this section, these capturing much of the moisture. Finally, the evidence given by the N curves, fig. 1 7 (assuming that the nucleation is not restored faster than it can be removed by the exhaus- tions), is much the same as that just detailed. The curves are often straighter than heretofore, and continually rise as the result of the in- creasing volume expansion; but the double inflection remains. In fact, the rise of all the upper curves is eventually at about the same rate as for air. In the latter case the scale is again too small to show the in- crease in the region of ions. At dp = 22 the distance effect for 1 and 6 meters is a decrease of 60 per cent, at ^ = 30 a decrease of 71 per cent, naturally nearly the same as above. Both the N and n curves show that while within the limits of observa- tion intense radiation increases the nucleation of dust-free air, this is not the case for weak ionization. Here the curves cross at about dp — 29 cm. Above this the presence of radiation must decrease the efficient nuclea- tion of non-energized dust-free air; below this it will increase it. At high pressure differences like dp = 41 cm., as used below, the variation of the efficient nucleation must therefore be a sensitive criterion for the variation of the ionization of air. Similarly the direct experiments of sec- tions 7 and 9 find their explanation in the manner already pointed out. 17. More rapid exhaustion. Apparatus and data. — The preceding set of experiments correspond to the particular exhaustion cock used. This was an inch plug gas cock. As almost the whole resistance to flow was encountered here, the advisability of enlarging it seemed evident. In fact, on opening the cock suddenly to an amount corresponding respectively to about 1/4, 1/2, 3/4, i/i,the nucleations obtained were about iVXio~ 3 = 4o, 100, 180, 350 for dp = ^i cm., showing that not even an approach to the limit is reached in the extreme case. There is a further objection in having the large resistance in the stopcock, as the coronas will vary with the degree of opening. 24 VAPOR NUCLEI AND IONS. Table io. — Nucleation of dust-free air, energized or not. New apparatus, i^-inch stopcock; connections of inch gas pipe. Barometer 76.23 cm. Walls of fog cham- ber 0.3 cm. thick. Three horizontal wet cloth partitions within chamber. dp. s. n x 10- 3 . N x io- 3 . A* l dp. S. »Xio J . iVxio- 3 . I. Ill (continued). X-rays. D=5ocm. from side. 44 »77 180 440 42 x 7 6 i75 400 35-2 5 9.o 240 45o 7 6 i75 400 30.8 5 9-3 258 440 7 6 i75 400 26.6 5 9.2 229 360 35 '7 6 160 300 22.8 3 8.o r 37 200 7 6 160 300 21.5 3 7-3 120 170 30.8 2 7 5 i37 235 20.8 6 7-2 100 138 26.1 2 6 62 9-7 19-5 5-i 39 52 24-3 7 . 2 •4 18.6 1 . 1 1 5 . 1 .2 18.3 ?o ?o ?o 25-3 26.1 1 2 4 6 I.-7 6.2 2.6 9-7 27.0 3 6 17 27-5 27.6 5 46 75 IV. X-rays. D = 1 2 cm. from side. 28.6 29-3 5 6 4 3 60 92 99 153 30- 5 7 I2 5 212 35 .0 x 7 7 160 300 18.9 2 2 2.7 3-7 35-o 30.5 2 7 8 5 160 136 300 230 20.0 21 .0 23.0 24-5 5 3 7 7 7 5 9 7 4 3 5 54 117 125 195 74 165 180 290 II. Rad ium tube 2 idded on side. Barome- 29-3 8 IO 5 290 490 35° 8 IO 7 320 610 tet 76 . 05 cm . (Mean values.) 43-7 5 IO 3 306 750 52.2 *9 4. 240 760 43-6 4.4 40 95 35-o 5 4 67 126 30.8 5 6 69 120 V. Dust-free air. Fog chamber with 4- 26.4 24.2 22.8 5 5 5 6 6 6 64 60 57 97 87 87 sheeted wet cloth. 2 to 3 cm. apart. Barometer 76.36 cm. 21 . 2 20.0 5 4 6 1 55 20 74 28 18.8 28.8 5- 1 19.0 27.0 3 4 19.0 19.6 •5 2.6 . 1 5-2 6.*9 25-3 25.8 2 1 20.4 4-9 37 5i 25.3 ? 21 .0 5-2 43 60 25.8 2 3 III. 3 C-rays. L ) = 50 cm. from side. 30.8 31.6 2 2 8 7 8 7 1 3 3 4 i 32.3 ( 6 ) 48.9 3 7-6 190 560 33- 1 6 7 7 43-6 4 8.o 216 530 34-o 9 7 8 g' BP corona steadily repeated, but not quite clear. *wr g. >wyg. 'gyog'. ! wog. 3 GBP. 8 wog'. * w P corona. 8 Vague. EFFICIENCY OF FOG CHAMBER. 25 Accordingly the inch stopcock was replaced by a ij-inch plug gas cock, retaining the same inch tubing. The resistance in this case was least at the stopcock, which accounts for the remarkable steadi- ness of the results obtained in successive exhaustions. Furthermore, the fog chamber was modified by inserting three wet cloth partitions, horizontal and about 5 cm. apart, as shown in fig. 3. The new form was not merely favorable to better saturation, but all con- vection currents were more effectually cut off. This was at first (though incorrectly) supposed to be the reason for the appearance of distribu- tions of nuclei within the chamber produced by the action of radium kept outside of it, a result already detailed above, but which here appeared for the first time in a very marked degree. In the later experiments four cloth partitions were inserted, as in fig. 4. The results are given in the successive parts, table 10, in the usual way, dust-free air being examined with and without the interference of external radiation. It is not probable that any errors are now intro- duced by the filter. In view of the intense ionization used, the usual cycle of measurable coronas is exceeded, and the resulting data become more and more fully relative, both on this account and because of the high pressure differences applied. Absolute values are, however, of little interest, and it is precisely from the relations obtained that conclusions are to be drawn. The quantities s, n, N, dp have been frequently defined. D is measured from the side of the apparatus, so that all action takes place through 0.3 cm. of glass. No aluminum screen was interposed. 18. Remarks on the s=curves. Non-energized air. — The present 5-curves for non-energized air, fig. 19, are in strong contrast with the preceding. Variability in the lapse of time has been nearly eliminated. The curves have moved bodily somewhat farther to the left into the region of smaller super saturations, as is shown by comparison with the dotted line in fig. 19 (taken from the preceding fig. 15), evidencing the increase of efficiency referred to. The curves of fig. 19 are in remarkable contrast with the preceding in two respects; they reach a definite asymptote at about dp = 35, which is retained until the pressure difference is so high that other complications step in. There is certainly an incre- ment after dp = 30, as proved by the change of coronas. Whether this limit is due to a cessation of further drop of temperature in the appara- tus, or whether a distinct group of colloidal air nuclei is in question, remains to be seen. The possibility of the continued succession of groups of this kind is not, a priori, improbable. 26 VAPOR NUCLEI AND IONS. 19. The same, continued. — The other characteristic, which was entirely unforeseen, is the relative absence in what has been called the ionized region (placed in the former curves, fig. 15, between about dp = 21 and dp = 26). The contrast is clearly shown in fig. 19. The only ex- planation which suggests itself to me is this, that what were supposed to be ions were really water nuclei due to the somewhat slower exhaus- tion of the preceding experiment. In other words, particles caught at the end of the somewhat less rapid expansion evaporated into the larger particles, leaving water nuclei behind. But apart from this, the definite trend with which the curve reaches the abscissa is characteristic, and, so far as coronas go, there seems to be an absence of nuclei below dp = 24. In other words, the fog limit has been raised, in spite of the general lower- ing of the necessary supersaturations throughout the curve. The new curve is almost wholly coronal. WOff' wPcor ZO ZZ 24 Z6 ZB 30 32 34 36 Fig. 19. — Improved fog chamber. Apertures of coronas (s) seen in dust-free air, energized or not, as stated, by radium or X-rays, at different exhaustions (dp). Table 10. Usually the same results are obtained in a pressure-increasing and in a pressure-decreasing series. There is an equally striking uniformity in the lapse of time. EFFICIENCY OF FOG CHAMBER. 27 20. The same, continued. Action of radium. — The action of radium is characterized by a slight rise of the asymptote over the preceding case and of displacement of the curve into smaller supersaturations. This was to be anticipated; nuclei are caught more easily, and a some- what smaller order of size and increased number is detected in conse- quence. The other distinctive feature is the smaller range of pressure differences within which practically all condensation occurs. In the preceding experiments this range lay between dp = 18 and 24, whereas in the present case the marked changes fall between dp = ig and 2 1 . There results another apparent rise of the fog limit corresponding to the coronal effects, though it is slight and uncertain. What is noteworthy is the greater steepness of the rising branch of the curve, already indicated. The question may be asked whether for a rigorously instantaneous exhaustion this part of the curve would become even more nearly vertical. One may note that the radium curve eventually passes through a maximum. Furthermore, while in its initial stages it lies very near the former X-ray curve (D = $o), the latter shows no tendency to reach a near asymptote. 21. The same. Action of X=rays. — Finally, the results with the X-rays are similar. High asymptotes, a tendency of the curve to lie within a region of lower supersaturations than in table 7, etc., are apparent. Nuclei smaller in size and larger in number have been caught more easily than heretofore. If we regard the present and the earlier curves for D = 5o, the height of asymptote in the former case is much above the one in the latter case, and moreover passes through the maximum (reached at dp = 26) as dp increases indefinitely. This cor- responds to the result for radium. It is therefore interesting that if the radiation is further intensified, as in the last case, where D = 1 2 cm. , the curve continues to rise apparently throughout higher ranges of pressure. The limits are not reached (dp continually increasing) until nuclei of a size equal in smallness to the order holding for the colloidal nuclei of air have become available for condensation. A final point deserving comment is the break of the X-ray curve near dp = 22. This is not only marked in both curves of table 10, but similarly apparent in table 7. Unfortunately, the cycle of coronas changes at this point (green-blue-purple corona); but apart from this it seems probable that with intense X-ray radiation, two suc- cessive groups of nuclei are in question ; the first group prominent below dp = 2 2 and the second above dp = 22. 22. Remarks on the w=curves. Non=energized air. — The w-curves, which assume that the nuclei are replaced sooner than they can be 28 VAPOR NUCLEI AND IONS. removed by exhaustion, bring out the points just made with better regard to the actual nucleation, seeing that the s -curves are specially favorable to small numbers of nuclei. As a rule the new nucleations obtained are nearly twice as large as the old results. It seems probable that the upper inflection, after passing the nearly straight part, is due f eeo 240 18 20 22 24 26 28 30 32 34 36 38 40 Fig. 20. — Improved fog chamber. Efficient nucleation (n) free air cor- responding to fig. 19. Table 10. to the waning condensing power of the apparatus, and that in the absence of this loss of efficiency the middle region of the doubly inflected curve would continue indefinitely. One may note, too, that both in case of tables 7 and 10 the parts in question are nearly parallel, or that the increment of nucleation corresponding to a given increment of super- EFFICIENCY OF FOG CHAMBER. 29 saturation is about the same. This applies even in the following cases, where dust-free air is energized by radium and by intense X-radiation, as far as the middle regions of the curve are concerned. 23. The same. Action of radium. — The small increment of pressure difference (say dp = 19 to 21) is here in sharp contrast to the gradual increase observed in the old results (say between dp = 19 and 25). The nucleations which eventually are not quite doubled in the new results, as compared with the older, are therefore at first enormously in excess. In fig. 20 the results of table 10 are given in full, the old results (table 7) in broken lines, and the same method is pursued in the other cases. 300 22 24 26 28 30 32 34 36 38 40 42 Fig. 2i. — Efficient nucleation (N) in dust-free air corresponding to fig. 19. Table 10. 24. The same. Action of X=rays. — Remarks to the same effect may be made with reference to the old and the new data for X-ray ionization when D = $o (broken and full lines in fig. 20). Initially the old curve is even below the present curve for radium, but it eventually intersects the new data for dust-free air at about dp = 3 5 . The mean nucleations have not been quite doubled. The maxima in cases of D = 50 and D = 1 2 lie above the interval of large variation for dust-free air. Finally, the breaks in the curves at dp = 22 are again quite apparent. 25. Remarks on the A r =curves. — With regard to the iV-curves, fig. 21, one need merely note that the maxima have in most cases been defi- nitely wiped out, and that the nucleation available for condensation, if it be supposed that removal of nuclei is more rapid than the reproduction, increases continually with the supersaturation up to the highest values. 30 VAPOR NUCLEI AND IONS. THE NUCLEATION OF FILTERED AIR IN THE LAPSE OF TIME. 26. Method. — In the above experiments with filtered air, it has been carefully pointed out that all the curves tend to reach asymptotes, the height of which depends on the other (larger) nuclei present. If, then, the ionization of dust-free air is an essentially variable quantity, as shown by the electrometer, the height of the asymptote in question should be correspondingly variable. Again, it was shown that external radiation (gamma-rays of radium for instance) are powerful nucleators, whether directly or secondarily. Hence external radiation of this type, if variable, might be reasonably sought for in a study of the nucleation of dust-free air. In relation to the ordinary intensities of radiation, this method of varying asymptotes is exceedingly sensitive. Whether, however, it will apply for cases of much weaker (cosmical) radiation must be left to experiment. Moreover, whether the variable ionization of air can be detected in the manner suggested, or whether it even exists in the fog chamber, is similarly a question requiring experimental solution. 27. Early data. — The method here in question was carried out in a series of experiments begun May 9 and continued until September 2, 1905. In the earlier part of the work it was shown that comparable results could only be expected in cases of excessively slow filtration, the filters for this purpose being 18 inches long, 2 inches or more in diameter, and containing cotton very tightly packed. It was necessary, moreover, to make observations after long intervals (12 to 24 hours) of waiting, the object being to allow all nuclei which might have passed through the filter time to decay. Apparent variations of considerable interest were obtained in this way ; but the work below will show them to have been untrustworthy. Neither changes of temperature nor of the barometer produced any effect. To additionally safeguard the work, two fog chambers were eventually installed side by side, drawn upon by the same exhaustion system identically in every way. It was then found that the internal partitions of wet cloth were essential, but it was nevertheless impossible to make both chambers agree even when the partitions were increased four- fold. Throughout the work, nucleations passing through maxima and minima within the limits of about 400,000 and 160,000 were obtained, apparently under trustworthy conditions. The data remained consistent even when the fog chambers were exchanged. Small variations of the rate of filtration were quite ineffective, showing that a limit has been reached. NUCLEATION IN LAPSE OF TIME. 31 28. Apparatus modified. — The tubing of the preceding apparatus, which consisted of inch plug cock and inch gas pipe, was now taken apart and a 1 J-inch brass plug cock was substituted. The advantage of this arrangement is easily seen, inasmuch as the whole resistance to flow is practically in the pipes. The passage-way in the stopcock being larger in area, the rate of exhaustion obtained is independent of slight differences in opening. As a result of this, all the variations noted above practically disap- peared, and the nucleations of dust-free air at a given pressure difference were found to be constant in the lapse of time within the limits of accuracy of the method. Data of this character are shown in the following table : Table 11. — Nucleation in the lapse of time. dp =30.6 cm. wo g' iVXlO -3 =227. nxio- 127; Date. Corona. s. Date. Corona. s. Sept. 5 1 wog 7-4 Sept. 12 wog 6 wog 7 4 13 wog 7 7 wog 7 4 14 wog 7.2 8 wog 7 4 15 wog 7 9 wog 7 4 16 wog 3 7-1 9 J wog 7 3 18 wog 7.2 10 wy'g 7 4 20 wog 11 wog? 21 wog 7.2 1 Apparatus with three sheets. 1 Apparatus with four sheets. 8 Apparatus with one sheet, 10 cm. above surface of water. Indeed, the succession of different fog chambers with 1,2, and 3 cloth partitions all gave the same results. The rate of filtration could be varied without effect within wide limits, and the same corona was obtained after a few minutes or after 24 hours of waiting. 29. Inferences. — It has, therefore, not been possible to detect changes of nucleation, either as the result of the concomitant changes of atmos- pheric ionization or as the possible result of some form of cosmical or at least external radiation. The consistency of the results obtained for radiation under widely different conditions may be especially pointed out. It does not follow, however, that other methods (not depending on the terminal asymptote) may not be more efficient. In how far this is the case will be shown in Chapter VI. 32 VAPOR NUCLEI AND IONS. SUMMARY OF THE RESULTS OF THE CHAPTER. 30. Distribution of ions within the fog chamber. — Within the fog chamber, the coronas due to the ions produced either by radium or by the X-rays acting axially from long distances, as a rule vary in aperture from one end of the long vessel to the other. Hence the distribution of nuclei is not uniform in the direction of the axis of the fog chamber. The reason for this is to be ascribed to secondary radia- tion. As interpreted from the nucleation produced within the fog chamber containing dust-free moist air, dense bodies (brass cap) produce more secondary radiation than rare bodies (glass end) under like con- ditions of primary radiation. If the primary radiation is relatively intense, rare bodies like glass may produce more secondary radiation than dense bodies (like the metals) under less intense primary radiation; or at least the decrement of primary radiation with the distance from the source predominates within the fog chamber. The secondary radiation is rapidly absorbed by the air, as the dimin- ishing coronas testify. 31. Minimum of efficient nucleation. — When ions and colloidal nuclei are in presence of each other (or in general any two groups of nuclei of different average size), the number of efficient nuclei for the case of a drop of pressure not too small passes through a well-defined minimum, if the number of ions continually increases from zero. This occurs, for instance, when a radium tube is moved from a long distance quite up to the fog chamber. The ions abstract more and more of the available moisture, until the colloidal nuclei are practically inactive (minimum), after which the nucleation increases again with the number of ions. 32. Persistent nuclei. — For the moderately intense X-ray bulb (5- inch spark) and a glass fog chamber several millimeters thick, persistent nuclei may be demonstrably produced when the distance from bulb to chamber increases to about 50 cm. They may even be produced through thin tin plate. The decrease with distance of the number generated within a given time is very rapid, and the number producible if the anticathode could touch the fog chamber would be enormous. In like manner the number increases at an accelerated rate with time of expos- ure as if the nuclei were themselves active. If we regard the ionized state of a gas as characterized by a kind of kinetic ionization pressure, we may further conceive the ionized gas to NUCLEATION IN LAPSE OF TIME. 33 be subject to something resembling saturation, whereby ions pass into persistent nuclei when the ionization pressure reaches a certain limit very much as a vapor condenses. 33. Dependence of efficiency of fog chamber on the size of the exhaust pipes. — If a (long) cylindrical fog chamber, whose contents are not much in excess of 6,000 c. cm., be exhausted into a vacuum chamber whose contents exceed 100,000 c. cm., through a passage-way 50 cm. in length and about 2.5 cm. in bore, many of the experiments showing the properties of ions and of the colloidal nuclei of dust-free air are con- sistently producible, provided the plug exhaustion stopcock (opened quite and as quickly as possible) has a wider bore than the pipes. With exhaust pipes of this size, however, the apparatus is as yet very far from its limits of efficiency, and the largest or terminal corona obtainable in case of the colloidal nucleation of dust-free air is, even at the highest exhaustion, the intermediate green-blue-purple type. This leaves the whole order of large coronas outstanding, though the order may be entered if the air is intensely ionized. It follows that efficiency actually stops short because of the smallness of the colloidal nuclei. The curves showing the nucleation entrapped in case of continually increasing exhaustion all have a common feature; there is an initial low branch of small variation attributable to ions; an intermediate branch showing steep ascent, due to colloidal nuclei; a final asymptotic branch, corresponding to the terminal corona and due to the loss of efficiency of the apparatus. The middle branch in question, or line of representative colloidal nuclei, if it appears at all, has almost the same slope and position (relative to the field as a whole) for all fog chambers, whether their efficiency be small or great. The fog chamber ceases to act almost abruptly; in other words, the terminal corona or largest corona producible as the drop of pressure continually increases, appears in any fog chamber as an almost sudden departure from the representa- tive lines. This corona is larger or smaller, the asymptote higher or lower, as the fog chamber is more or less efficient. 34. Invariable character of colloidal nucleation in the lapse of time. — The terminal corona (in a given type of apparatus), if not too small (Chapter VI), is a fixture as to size in the lapse of time (months). It does not vary appreciably with the departure from normal ionization, to which atmospheric air is incident. Probably the number of available colloidal nuclei is too enormous, as compared with the small variations in number of the relatively large ions, to admit of an observable effect. CHAPTER II. LATER AND FINAL STAGES IN THE EFFICIENCY OF THE FOG CHAMBER DUE TO A GRADUAL INCREASE IN THE BORE OF THE EXHAUST PIPES. CONNECTING PIPES NOT LARGER THAN 1.5 INCHES IN DIAMETER. 35. Introductory. — It will be the object of the present chapter to further determine to what an extent a perfectly efficient apparatus (i. e., one in which the cooling and the moisture precipitated per cubic centimeter actually accords with the drop in pressure) may be ap- proached by gradually enlarging the bore of the escape pipes between the vacuum chamber and the fog chamber. One method of doing this will consist in finding whether the distribution curves eventually reach a fixed limit and a limit maintained for larger as well as smaller nuclei; another by comparing the final results with corresponding data deduced for Wilson 's piston apparatus. In treating the large numbers of observations which follow, it will be convenient to refer the nucleations (as above) to the drop of pressure observed under apparently isothermal conditions at the fog chamber. This method of comparison is sufficient if the same fog and vacuum chambers are used throughout, as has actually been done in the sequel. The exhaust cock is always to be closed immediately after exhaustion. In the present experiments the volume ratio of the two vessels was about v/V = 0.064, where v is the volume of the fog chamber and V that of the vacuum chamber. It was customary, as just stated, to open the stopcock between the fog chamber and the vacuum chamber as rapidly as possible and then to close it at once ; dp therefore denotes the drop of pressure read off on the mercury gage at the fog chamber after sufficient waiting. As above, s is the angular diameter of the coronas on a goniometer radius of 30 cm., when the eye and the source of light are respectively 40 cm. and 250 cm. on opposite sides of the fog chamber. On the basis of these data n is computed from the quantity* of water precipitated per cubic centimeter and the value s, in the way shown elsewhere. f It is always assumed that the nuclei are more rapidly reproduced than removed by the exhaustion, or (without hypothesis) from another point of view, ♦Wilson: Phil. Trans. Roy. Soc. London, vol. 189, 1897, p. 298. f Smithsonian Contributions, vol. xxxiv, No. 1651, 1905, chap. viii. 34 EFFICIENCY OF FOG CHAMBER. 35 that the nucleation of the exhausted fog chamber is specified. Otherwise it would be necessary to multiply by the ratio of volumes after and before exhaustion. Diagrams of the apparatus used have already been given in Chapter II (figs, i to 4), and further examples will be shown of the details of the more perfected forms below (figs. 26, 34, 64). In those paragraphs, moreover, I will treat the corrections to be applied (which are much larger than were anticipated) in order to pass from the observed appar- ent dp to the true values. These must be computed from the volume ratio of the two chambers and their respective initial and their final isothermal pressures when in communication. 36. Examples of data for 1-inch connecting pipes. — These were about 18 inches or more in length and the stopcock inserted was i\ inches in bore, as specified in the preceding chapter. The plug was as usual floated in paraffin oil, absolutely preventing the influx of atmospheric air, though some leakage (which is generally harmless) from the fog chamber to the vacuum chamber could not be prevented. The fog chamber in these exhaustions had a rather larger charge of water than usual, a circumstance to which the higher fog limits may possibly be due, since these vary with the volume ratio. Table 12. — Pipes 1 inch, cock i\ inches. Non-energized dust-free air. October 24. dp. s. n xio- 3 . 30.6 7-4 139 28.9 5-8 74 27 2 4.2 26 25-7 i-5 i-9 24.6 The results are to be shown below (fig. 22, p. 37) in comparison with the later values and may be passed over here. 37. Data for pipes 1 . 5 inches in diameter. — Fog chamber and vacuum chamber were now connected by brass gas pipes ij inches in diameter and 2 feet long, and the plug stopcock interposed was 2 inches in bore. In the case of table 1 3 the disposition of apparatus was such as to require the length of piping specified with an elbow. The results are given in the usual way and show a distinct gain in efficiency over the earlier set in table 1 3 , proving that greater width of pipe is still an advantage. 36 VAPOR NUCLEI AND IONS. Table 13. — Nucleation of dust-free air, energized or not, as stated, at different super- saturations. Plug gas cock, 2 inches; piping i£ inches in bore, 2 feet long. dp. ttX io" I. Non-energized air. cm. cm. 30-7 x 7.2 28.2 59 26.3 3-7 24.4 2.7 22.4 2-3 20.3 .0 22.0 1-7 24.7 2.7 27.6 6.1 29.9 2 7-3 34-2 s 8.i 151 77 19.7 6.8 3-3 .0 2.1 6.8 80 135 190 Non-energized air. Exper- iments. .8 8.4 • 7 f 4 IO.O .8 I 6 8-9 .6 { 4 io-7 •5 \ 5 9.o 190 270 220 270 220 II. Non-energized air. 42.1 e 9-3 34-5 7 io.5 32.1 8 9. 2 30.0 B 8.o 27.9 6.1 26.0 3-8 24-3 2.7 22.5 2.0 21. 1 ?I.O 173 !.o 18.6 !.o 250 318 262 150 80 20 6. 2.3 r .0 .0 nX io" II (continued). Non- energized air. 19.9 ?o.8 20.4 I .2 21 .O 1.6 0.6 I . 2 i-7 Later. 43-5 hi. 2 47-9 7 II.2 43-8 7 io.8 39-1 8 IO.I 34-9 10 8.9 30.8 n 7-5 354 372 338 286 192 144 III. Air energized by radi- um, on side of glass fog chamber. 21.7 6.0 20.9 5-7 20. 2 5-2 18.6 1.0 16.5 .0 >i8.7 I .2 18.8 i-9 21 .0 ?5-7 20.0 5-4 26.4 6.8 24.9 6.8 23-5 6.8 22.0 6.7 20.0 5-4 29.0 6-5 33-2 6.0 38.9 5-9 43-6 5-7 67 56 42 .8 .0 ?1.2 2.0 56 47 104 100 96 89 47 99 88 9i 85 dp. n X io~ 3 . IV. Air energized by X- rays 12 . D = 20 cm. from side of fog chamber to anticathode. 34-7 l3 i2. 3 31.0 13 I2.3 26.3 13 i2.3 22.0 14 io.6 17.6 1.4 16.5 .0 17.6 2. 1 19.6 15 8.i 19.0 4.4 400 380 340 207 I o 2 I4O 24 V. X-rays. Z?=5ocm. 32. 35< 26. 22, 17 17' 19 20. I "11.4 16 n.3 6 14 io.7 1 17 8.6 4 ?r 7 ?o 2.2 6.1 310 320 233 145 o o 2.7 64 X-rays. D = 20 cm. after long waiting. 34-6 g' B p, repeated. 2 gyobg. s w p cor. 'wvog, after waiting one or more hours. 8 w' P cor after waiting but a few (io to 15) minutes. 6 wcg. 7 wog. 8 wrg. 9 g' B P. 10 wP cor. "wycor. "Air non-energized, ftp— 348; 8=9.6; and by X-rays. 1J wyg r large, "wrg. 16 w c g. "wobg. 17 w P cor. 18 gyobg. Long waiting before exhaustion (to remove water nuclei) is often essential. In addition to the dust-free gas, air energized by radium (io mg. of a weak sample, io,oooX, sealed in aluminum and acting from without) and by the X-rays is treated for comparison. EFFICIENCY OF FOG CHAMBER. 37 400 360 280 Z40 Fig. 22 — Efficient nucleation in) found in dust-free air energized or not by radium or the X-rays, as stated, from different distances (D), at different exhaustions (dp). Fog chamber with pipes ij-inch bore, 2 feet long. Table 13. 38 VAPOR NUCLEI AND IONS. These results, together with the data of table 12, are shown in fig. 22. For the air curve a marked advance beyond the results of table 1 2 is at once apparent, but it is noteworthy that the slopes of both curves are in the main the same, in spite of the fact that the former rises to but one- half or one- third of the height of the new curve. The efficiency of both fog chambers therefore terminates abruptly in a final or fixed corona or (in the chart) in an asymptote which would be horizontal, if it were not necessary to assume that the quantity of water, m, precipitated per cubic centimeter increases with the drop in pressure. In the different series, irregularities are often apparent which are difficult to explain, though at low values they may be due to cosmical radiation. Thus the second series for non-energized air rises more abruptly than the first and third, results which are possibly associated with the presence of water nuclei and referable to different degrees of leakage from fog chamber to vacuum chamber; but it is also possible that unequally rapid closing of the stop- cock after exhaustion may have left an impression. The results for radium and the X-rays will reappear under better conditions presently, and may then be discussed. One may note that for weak radiation the graphs intersect the graph for non-energized air. 38. Continued. Shorter pipes. — The next step consisted in a remodel- ing of the apparatus, whereby shorter connecting pipes would suffice. Two 6-inch nipples, i£ inches in diameter and containing a 2 -inch plug stopcock, now joined the fog and vacuum chambers. The results are shown in table 14 and in fig. 23. The fact that the main resistance is still encountered in the gas cock is proved by the need of stops to insure quick and complete opening. Whenever the cock is nearly but not quite opened, however suddenly, low coronas are observed. When this was remedied, fixed coronas thereafter corresponded to each dp. The table contains the usual series of comparative results for air energized by radium and the X-rays. As compared with the preceding data there is a small shifting of the graph toward lower pressures, while the slopes in the representative parts of the curves are practically the same. The terminal corona, however, and the corresponding asymptote, is markedly raised. Nevertheless, though the large green-blue-purple corona now appears in the case of strongly energized air, it has not yet been reached in the non-energized cases. The limit of efficiency is therefore not yet in sight when a fog chamber of about ^ = 6,000 c. cm. empties into a vacuum chamber of about V = 100,000 c. cm. through 30 cm. of piping about 4 cm. in diameter, with a corresponding plug stopcock suddenly opened between stops to arrest the motion. EFFICIENCY OF FOG CHAMBER. 39 Table 14. — Nucleation of dust-free air, energized or not, at different supersaturations. Plug gas cock, 2 inches ; piping 1 £ inches in diameter, 2x6 inches long. Long intervals between observations. dp. Corona. n Xio" I. Non-energized air. 1 39.8 12.8 39-8 130 35-1 13.0 30.8 10.4 28.6 7-4 27.9 5-7 26.0 2.4 26.4 3-6 gyobg wybg wyg wr g wrg w B P cor cor II. Non-energized air. 1 243 1.5 24.4 1.8 26.2 2.6 27.3 5.o 28.0 6.4 293 7.6 295 8.0 30.9 10. 30.4 8-4 31.2 9-i 350 "•5 cor cor cor cor wrg gBP gBP + w r o g wcg wrg wyg" III. Non-energized air. 31.0 9-8 31.0 10.3 310 2 io.4 35.3 12.0 351 12.0 39-4 12.0 44.1 12.5 41-3 12.5 38.6 12.0 307 10.2 w r o g wr og wr og wybg wybg wybg gyo gyo wybg wrg 460 420 400 260 108 69 45 17 2 6 46 92 139 147 290 230 260 400 290 290 290 400 400 420 490 470 420 290 IV. X-rays at D = 20 cm. from side. 30.9 130 35-4 130 33-0 13.0 31-5 13.0 29. 2 130 gBP gBP gBP gBP gBP 460 480 470 460 430 dp. Corona. WXIO" IV (continued). X-rays at Z?= 20 cm. from side. 27.2 12.7 25.6 12.7 24.1 12.2 22.8 11. 9 29.4 12.3 27.1 Vague 20.8 7-7 23-4 11. 7 22.9 11. 4 21.4 8.4 20.2 4.6 18.7 .0 18.3 .0 gyo gyo wyg' wog gBP gBP wog wog gBP cor 380 360 320 250 440 410 120 250 250 120 30 o o IV. X-rays. D = 10 cm. 351 14 40- 3 37-9 33-6 31-6 30.0 28.1 26.7 25.0 13 5 23.6 12 6 22.3 10 3 21.2 7 1 20.1 4 2 18.7 1 3 3gB gB gB gB gB gB gB gB 4 gB gyo wog wog cor 480 5io 500 470 460 450 430 410 400 320 250 100 21 1-3 V. Air 8 energized by radium, D side. o, on 30- 8 6.1 35-2 6.2 40.0 59 27.6 6.0 26.2 6.1 24- 3 6-3 22.9 6-5 21.6 6.1 21 .0 4-7 19.7 2. 1 19.2 .0 cor cor cor cor cor cor cor cor cor cor 87 95 92 79 78 79 85 68 33 2. 1 In neither series I or II were stops provided for the exhaust cock. This is the case in series III, et seq. 'Filter open 90 without danger. Evaporated corona shows *=» 7.6. •Coronas at first stone-blue and irregular, soon becoming gBP. *CleargB P at once. •For air, «= 9.9 w o g. n X io- 3 = 290. Large coronas fall out, leaving a smaller one behind. 4o VAPOR NUCLEI AND IONS. Several interesting features are seen in the radium and X-ray curves in fig. 23, to which special attention may be called. It will be observed that the curves for distances D = io and 20 cm. of the anticathode from the sides of the fog chamber begin and terminate in nearly coincident curves and coronas, but that they lie at some distance apart in their middle regions. Hence available* nuclei are apparently not more numerous eventually in one case than in the other, but they are larger (virtually) in the middle regions for the stronger radiations; in other words, the asymptote is more quickly approached for the stronger radiation. 300 ZOO 20 Z5 30 35 40 Fig. 23. — Efficient nucleations (n) in dust-free air, energized or not, as specified at different exhaustions (dp). Table 14. Again, the terminal corona in case of the data for X-rays is reached long after the final coronas for the weak radiation from radium (10 mg. io,oooX); i.e., the number of nuclei (ions) in the case of the weak radiation ceases to increase in a region of lower exhaustion, apparently, than the one observed for the more intense radiation. Hence in case of the latter, finer gradations of nuclei appear to occur, as if the radiation shattered certain of the larger ions. There is, however, another point of *It will appear below that for nuclei, large or small, the limit is reached with the given number of nuclei per cubic centimeter, corresponding to the g B P corona. Nuclei in excess of this, large or small, are inefficient. EFFICIENCY OF FOG CHAMBER. 41 view to which some probability may be assigned. In the case of intense radiation relatively large and possibly persistent nuclei are produced in greater numbers, and these may capture much of the moisture and prevent condensation on the smaller ions, until proportionately larger exhaustions have been applied. Examples of these occurrences will be given in the next paragraph. CONNECTING PIPES TWO INCHES IN DIAMETER. 39. Remarks on the method. — The connection between fog cham- ber and vacuum chamber was now further enlarged by using two nipples 6 inches long and 2 inches in diameter, on either side of a plug stopcock of 2\ inches bore. The latter was chosen, since much of the resist- ance was heretofore encountered in this place. Usually observations were taken at long intervals of time apart, to allow for the dissipation of water nuclei due to the evaporation of very small fog particles. Moreover, a considerable period (three days) elapsed before definite results could be recorded. Something of a nuclei -producing character is usually present some time after assembling the parts of the apparatus. The coronas under otherwise like circumstance gradually increase to a maximum. Moreover, in the absence of all leakage inward from without, periodicity of coronal diameter is often in evidence in the successive exhaustions and especially active in case of the very large coronas. The best results appear after long waiting before each observation, whence it follows that evaporation from the heavy paraffin oil used for tightening the stopcocks, leakage of nuclei through the filter or through unknown channels is quite ineffective. Waiting was finally made needless by first exhausting the partially exhausted fog chamber before each definite observation. The effective drop of pressure is then small and only the large nuclei are caught and the fog particles do not evaporate appreciably. The filter, moreover, in all tests gives evidence of entire trustworthiness. It was customary to open and close the plug cock quickly, between stops, a spring opening device having been put in action. In no case was there danger of under-saturation, the wet cloth linings being but 8 or 10 cm. apart, above and below. In the highest exhaustions used, ^ = 44 cm. about, the volume expansion is fully 2.5, so that 60 per cent of the air must thereafter be readmitted freshly through the filter. Naturally in this slow passage it would be not only denucleated but deionized, if there were not fresh sources of penetrating radiation always in action for the reproduction of both types of nuclei. In general the apparatus seems to have behaved faultlessly. Such irregularities as remain must then in large measure have been introduced by the fact that the drop of pressure 42 VAPOR NUCLEI AND IONS. was observed at the fog chamber when isothermal conditions had been reestablished. It should have been computed from the initial pressures of both chambers and their common isothermal pressure when in com- munication, in the way shown in the next section. The present results, however, suffice for comparison with the preceding and following work, and it is probable that limiting conditions of efficiency have already been reached. That this is actually the case will also be shown in the next section. 40. Data for pipes 2 inches in diameter, 12 inches long. — The results obtained are fully given in table 15. It appears that coronas, even above the large green-blue-purple type, appeared at the high exhaustions, but this is not quite certain. Incidentally some data on the effect of X-rays from different distances are included in the tables (part VI et seq.). The usual difficulty of an inconstant X-ray bulb appears. Moreover, certain peculiar results on the distribution of the nuclei within the fog chamber in case of symmet- rical exposure to weak radium are here again noticed (part V), as in Chapter I, section 4. The results are plotted in fig. 24 and smooth curves obtained therefrom are shown in fig. 25. Table 15. — Air nucleation at different supersaturations. Piping 2 inches in diameter, 12 inches long; plug gas cock, 2$ inches in bore. Wait 30 minutes between observa- tions. Angular diameter of coronas, 2 sin -1 s/60. Eye 40 cm., lamp 250 cm., on opposite sides of fog chamber, dp observed isothermally at the fog chamber. Coro- na. nx I. Non-energized dust- free air. 35.4 14 gBP 35-3 14 gBP 40.6 «4 gBP 44.4 14 'gBP 41.4 14 gBP 36.6 14 gBP 33-0 14 gBP 3* 1 gyo 29-5 10.3 wr'g 27.7 6.4 wr'g 29-3 10.4 wr g 28.0 ^6.7 wrg 26. 2 %2.6 Fcor fear 26.0 .'3-i 24.0 Ti O £cor 27.2 hi wy'g 480 480 510 540 520 490 470 410 250 91 250 104 6. 9- 120 Coro- na. nx Non-energized dust- free air — Cont'd. 28.0 7-7 wPcor 28.9 10.6 wrog 30.8 12.5 gyo 31.8 12.8 gyo 330 130 gBP 170 280 410 420 470 II. Air energized by X- rays; D= 10 cm. 33-3 35-0 39-8 3i-7 26.9 24-5 23-3 21 .0 ( 2 ) (*) ( 3 ) 13-5 b'BP b'BP b'BP n-5 b'BP 11. 7 gBP 13 wo bg 470 480 510 460 420 390 380 230 Coro- na. nx II. Air energized by X-rays; D = 10 cm. — Cont'd. 20.4 9.0 wcg 21 .O 11. 4 w obg 19.7 6.2 wr bg 19. I 4.1 cor 18.4 i-3 cor I7.4 .0 170 230 190 19 I o Continued, next day; D=* 10 cm. 18 18 18, 20. 20. 21 . 5 4 o 5 9 1.6 cor 3 6.8 wrg 9 9-3 wcg' 9 gyo o o 2 87 l80 330 1 Stone- blue colors, above gBP. 2 Irregular. Shows strong purple band just above water. Coronas diffuse. 8 Stone- blue. ♦Left over night seems to diminish rainlike condensation. EFFICIENCY OF FOG CHAMBER. Tabi^e 15. — Continued. 43 dp. Coro- na. nX III. Air energized by X- rays; D=6o cm. 175 19. 1 20.3 21 .0 22.5 22.8 23 -5 24.8 26.6 26.6 30.9 35-8 0.0 1.7 6.1 8.7 11. 7 12.2 l i3.o cor wr g wc g wog' wog' gyo gyo gyo gyo gyo gyo 2 65 180 250 270 340 350 370 370 410 440 IV. Air energized by X- rays; D=*6oo. 40.0 8.7 36.0 7-3 cor 3i-7 6.1 cor 27.2 5-2 cor 24.4 5-2 cor 23.0 5-2 cor 21.0 50 cor 19.4 .0 19.4 .0 18. 1 .0 20.3 3-i cor 21 .0 4.8 cor 23.0 5-7 cor 26.3 5-4 cor 310 5-8 cor 35-3 2 6.9 cor 35-4 6.8 cor 39-8 8.1 w r 44-3 10. w 44-4 S IO.O w 250 150 90 52 50 46 39 o o o 8 35 60 57 76 130 125 205 322 322 44-3 40.2 35-9 33-5 30.8 27.0 24.4 23.2 20.9 20.5 19.4 18.3 dp. Coro- na. nX io- s V. Air energized by radi- um (10 mg. 10,000 X); Z? = o, on side. 4 9-7 w o' 7-5 w r* 6.6 w 6.6 w r 6.5 w r 6.1 w r 5-8 w r 6.2 cor 5-8 cor 4-7 cor 1.3 cor .0 305 153 "5 in 103 79 436 74 60 32 1-3 .0 VI. Air energized by X- rays. '22.9 '22.9 22.9 '22.9 5-5 7-3 7-7 97 wr g wyg wlBP wcg no 140 200 VII. Air energized by X- rays; D= 200 cm. 19.0 2. 1 cor 20.3 6.8 wr g 20.9 7-4 wg 22.8 7.5 !gBP 25-9 7-5 gBP 3i-i u 7 ' 4 wyg 304 "7-4 wog 35-5 u 7 ' 4 wog 35-i "7-4 wog 40.0 74 wog 43-9 7.0 wrog 2 85 105 125 135 135 130 135 135 145 135 9P. Coro- na. n x VIII. Air energized by X- rays; Z?=6oo cm.; re- peated for contrast. 43-9 46.8 48.8 10.7 10.6 10.5 wog wog wog 300 310 320 IX. Air energized by X- rays; D= 100 cm. 44.4 7 6 w|BP 41.9 7 6 w|BP 36.2 7 6 gBP 31.8 7 6 gBP 25 -4 lOg 8 wcg 7 26.5 9 2 wrg 24.0 9 3 wcg 21.6 8 2 w|BP 20.8 7 4 wyg 20.3 5 5 cor 19.7 u 4 1 cor 28.0 8 8 wcg' 316 7 8 g|BP 180 170 160 150 200 205 195 130 105 50 19 215 170 X. Air energized by X-rays; D= 50 cm. from end. 30.8 8.8 wrg 35.6 8.7 wPcor 40.1 8.2 wPcor 230 220 205 XI. The same; #=50 cm. Bulb more efficient. 31.6 9-8 wcg 27.0 IO.8 wrog 25.2 IO.9 wrog 22. 2 IO. O wcg 20.3 6.1 wrg 35-1 9.2 wcg 4O.O 8.3 wPcor 44.O 7-8 w|BP 48.7 7-7 g , BP 32.2 10.4 wrog 245 235 225 200 65 240 205 180 190 260 1 As intensity of ionization increases there is less shattering. 2 Next day. Fog chamber absolutely tight. Coronas all blurred. 3 The final use in response to new conditions. 4 Coronas blurred and larger at the glass end of fog chamber. •D = 6oo cm. ®D = 3,000 cm. 7 D= 150 cm. *D= 75 cm. 9 Long waiting (£ hour to 2 hours) without effect. 10 Passing, as usual, through a maximum. 11 Fog limit below 20. 44 VAPOR NUCLEI AND IONS. 41 . The same, continued. X=ray bulb inclosed in lead . — In the experi- ments detailed the X-ray bulb was not inclosed, so that secondary radiation issued from the whole surrounding environment. In the data given in table 17 the bulb is surrounded by a wide lead box, containing a window 7.5 cm. in diameter fronting the fog chamber. The window was in a removable side or lid of the box. This was kept at a long distance (D = 6oo cm.) in order to furnish weaker ionization than was specified in table 15. h' 9 500 400 300 200 100 20 25 30 35 40 Fig. 24. — Efficient nucleations in dust-free air, energized or not by radium or the X-rays from different distances (D) and different exhaustions (dp). X-ray bulb not within lead case, unless stated. In the first part of table 16 the exhaustion, dp = 27 cm., lies below the fog limit of dust-free air. The low degree of ionization is thus shown under different conditions. The second series contains a number of successive exhaustions for an increasing dp, with the radiation passing through a thin tin plate. Parts III and IV of the table are similar DISTRIBUTIONS OF NUCLEI. 45 series, with the window open (tin plate removed), the last series being a detail for the initial part of the curve, which is of especial interest. The results are also shown in fig. 24, where their relation to the lowest curve, D = 6oo cm., for the open or non-inclosed X-ray bulb becomes evident. In each of these curves there is a definite horizontal branch within which the ions predominate. Beyond this lies the rise due to the colloidal nuclei, but the curves, and particularly the asymptotes, are depressed in both cases. Table 16. — Nucleations at different exhaustions. 2^-inch cock, 2-inch piping, 12 inches long. X-ray bulb in lead case at D = 6oo cm. from glass fog chamber. Aperture in case 7.5 cm. in diameter. Lead box. s. Cor. 1 dp. n Xio~ 3 . I. Window open 3-i .0 cor cor 20.5 20.5 8-4 .0 Window closed with lead. Window open 3-4 2.9 cor 20.5 20.5 11. 7 6.8 Window closed with tin 1 . cor Front lead side off 2 3-5 cor 20.5 13.0 II. Closed with tin plate. . . .0 cor 18.6 .0 30 cor 20.6 7-6 2.7 cor 22.3 6.4 2.8 cor 23-7 7.0 4.1 cor 26.8 24-3 12.0 wybg 31.2 380 gBP 34-7 480 III. Tin plate off 3 wy b g wog 33-8 310 400 305 5-2 w r 28.0 55 4.2 cor 26.6 26 4.1 cor 24- 5 23 4.2 cor 22.5 24 3-4 cor 20.7 12 2.6 cor 20.0 5-2 Tin plate on i-7 2-5 3-8 cor 20.0 i-7 4.8 17 IV. Tin plate off cor 20.0 cor 21.5 4.2 cor 22.0 24 4.0 cor 22.7 20 4.0 cor 24.0 21 1 Open side over i foot square. 2 Lead sheet o. 12 cm. thick; tin plate 0.03 cm. thick. 3 After long waiting. 42. Discussion. — It will now be in place to make a more complete survey of the manifestations of nucleation in dust-free air under any conditions. Figs. 24 and 25 are useful for this purpose. With regard to the method I may recall that the difficulties were enhanced by the need of providing rapid exhaustions in case of a fog chamber large enough to measure the coronas and to show the sequence of axial colors. 46 VAPOR NUCLEI AND IONS. The earliest results (Chapter I, section 12) scarcely captured the maximum of 30,000 nuclei under the extreme exhaustion corresponding to the limiting asymptote (marked "preliminary" in fig. 24). In proportion as the apparatus was perfected, this asymptote was raised, step by step, until nucleations of at least 500,000 were leftinthe exhausted fog chamber. Probably the asymptote, or more accurately the terminal corona, obtained in any given type of apparatus, indicates that the limit of condensing 30 c 35 DECR. SIZE—** Fig. 2 5 — Smooth'curve from fig. 24. Nucleation (AT) of dust-free air, in thousands of nuclei per cubic centimeter, energized or not, as stated, by weak radium, or by the X-rays, at different distances {D in cm.). The letters attached to the curves indicate the characteristic colors of the inner field (or its margin) of the largest coronas, the order of colors being R, O, Y, G, B. The abscissas show the observed isothermal drop of pressure by which the supersaturation is produced and the nucleations applied rela- tively to the given apparatus. The line marked "air" shows the behavior of dust-free non-energized air terminating in the large green coronas. In a perfect apparatus supersaturation would increase enormously beyond a; the droplets would be at freez- ing beyond b. In the given apparatus efficiency probably ceases beyond c when damp air is the medium. power has been reached. Between the fog limit and the terminal corona the graph rises in a straight upward sweep; and it is remarkable that the rates at which the nucleation decrease with the drop of pressure, or better with the volume expansion, are about the same between fog limit and asymptote, no matter whether they lie within the region of high supersaturation characteristic of the colloidal nuclei or in the region of relatively low supersaturation characteristic of ionized air. Moreover, DISTRIBUTIONS OF NUCLEI. 47 they are the same in any apparatus, within the limits of its efficiency. In fact, even the curve which I have endeavored to reduce to the same scale from Wilson's disk colors for non-energized dust-free air (apart from differences in the meaning of dp to be explained in the next sec- tion) lies in the same region of pressure difference, and shows a slope quite in keeping with the other curves.* For an imperfect apparatus, on the other hand, the slope terminates abruptly at a lower asymptote and in a region of higher exhaustion. It will be convenient to refer to the nuclei belonging to the slopes specified as representative nuclei. To obtain the curve for the colloidal nuclei of dust-free air, it is often necessary to wait a week or more, until unknown internal sources of nucleation have spent themselves. Moreover, the isolated observations made with advantage ought to be made hours apart, for it is with the exhaustion that the release of spurious internal nuclei occurs, suggesting that water nuclei due to the evaporation of small fog particles are possibly in question. The air curve passes through the first two cycles of large coronas, terminating beyond the highest green-blue-purple, actually in the stone-blue or higher coronas, the first of the series producible in the fog chamber by any means whatever. It corresponds with the opaque of the steam jet, beyond which, however, there exists another cycle of yellows, to my knowledge beyond the reach of the fog chamber. It is not probable that the medium is as yet quite optically inactive after the bluish coronas specified, but merely appears so in small thicknesses. If we turn from the curve of non-energized air to the case of weak ionization such as is produced by 10 mg. of impure radium (i 0,000 X) or by the X-rays from a distance of 6 meters, we may note in the first place that the coronal fog limit has enormously decreased as compared with its position in dust-free non-energized air. The nuclei are therefore throughout correspondingly larger in size. The nucleation soon rises with an approach to the characteristic slope referred to, showing that the representative ions lie within a relatively narrow range of sizes, the curve of distribution terminating in a maximum. With increasing exhaustion the nucleation first decreases to a minimum; hence to all appearances the ions are now partially destroyed or else materially reduced in size at low pressure, so as to fail of capture even in the higher exhaustions. Beyond the minimum the final and most interesting stage of variation characteristic of weak ionization may be observed. The nucleation rapidly rises again, eventually to approach the asymptote of dust-free air. In explanation of this phenomenon, we may agree that ♦The coincidence between my interpretation of Wilson's disk colors and my own data holds only when the drop of pressure is observed at the fog chamber. »See section 48. 48 VAPOR NUCLEI AND IONS. further nuclei are destroyed or rendered relatively inefficient by the low pressure until there is a sufficient excess of supersaturation to actually capture the colloidal nuclei of dust-free air more and more fully in the presence of the ions (few in number and small in size) remaining. Pos- sibly the relative number of ions and colloidal nuclei is itself a suffi- cient reason; but it since seems clear that an apparatus which ceases to produce condensation on colloidal nuclei much below dp = 35 cm. can not be expected to regain efficiency between ^ = 35 and dp = 45 cm., unless there is destruction of whatever has been holding down the effective nucleation to low numbers. As the ions decrease either in size or number, there is a reappearance of the normal colloidal nucleation of dust-free air. Its asymptote is less and less depressed. Under moderate ionization, such as is obtained from the X-ray bulb at a distance of 1 to 2 meters, the fog limit is definitely reduced, showing the presence of an order of larger nuclei; the asymptotes are correspond- ingly higher and they are reached later (i. e., at higher exhaustions), indicating the presence of smaller nuclei than in the preceding cases, as well as larger nuclei, unless the increased presence of the latter retards condensation on the former. The range of sizes within which the repre- sentative nucleations lie is definitely extended in both directions, but particularly on the sides of the smaller nuclei. The maximum and the minimum are flattened, though destruction of ions still occurs at low pressure. The new feature of these curves is their failure to rise from the mini- mum toward the asymptote of dust-free air. It follows that the ions are now present in too large a number, even relatively to their reduced sizes at high exhaustion (if this obtains), to leave an excess of super- saturation sufficient to catch the colloidal air nuclei; or, from the other point of view, the kinetic ionization pressure is too strong to admit of sufficient rupture of the nuclei even at the lowest pressures applied. (Section 43.) In the comparison of both series of experiments, the occurrence of the intersection of the curve is noteworthy. In other words, the nuclea- tion, apparently produced at high exhaustion by an X-ray bulb at 6 meters from the fog chamber, may be much larger than the nucleation produced from 200, 100, or even 50 cm., due (as we have supposed) to the reappearing efficiency of the colloidal nuclei in the former case. Similar complications surround the distance effects at other pressures; but under conditions of sufficiently moderate exhaustion, i. e., below the fog limit of filtered air, the distance effects between 50 and 600 cm. remains disproportionately small. When the ionization is relatively intense, as when the anticathode is from 10 to 60 cm. from the thinner side of the glass fog chamber, the DISTRIBUTIONS OF NUCLEI. 49 fog limit has decreased somewhat further, showing increased abundance of the larger nuclei; but the slope closely resembles the corresponding case for dust-free air. There is no maximum, but a terminal asymptote, or better, a terminal corona of fixed aperture. This corona, which is here the largest obtainable in the fog chamber, is even at the highest ionizations reached at supersaturations below the coronal fog limits for dust-free air, and long before the apparatus has ceased to function. It is impossible to state whether colloidal nuclei are simultaneously present but inactive or whether their inefficiency is due to actual absence. It thus becomes a problem of great theoretical importance in its bearing on this subject to determine how the precipitated super saturation is distributed among groups of nuclei of different sizes. It is finally to be observed that the terminal corona is the same, both for the ionized and the non-ionized state of the gas. The maximum number of nuclei which can be caught per cubic centimeter seems, there- fore, to be a constant for the apparatus and the medium inclosed, and to be independent of the size of the nuclei, whether large like the ions or small like the colloidal nuclei. In the above apparatus, with an air- water medium it is difficult to pass beyond the large green-blue-purple corona. 43. Radiant fields. — The occurrence of a succession of groups of colloidal nuclei of (let us say for simplicity) continuously decreasing size and continuously increasing number, is suggestive; for each group is essential in the given natural but otherwise unknown environment. They are at once restored if withdrawn. These conditions may be imi- tated or varied artificially by approaching the radium tube at different distances from the fog chamber, in which case the efficient nucleation will for weak radiation usually decrease, as the intervening distances are smaller. Furthermore, in the presence of radium the character of the phenom- enon is the same, except that the nuclei are throughout larger. With- drawn by precipitation they are at once restored. They are an essential part of air in the new (radiant) environment and the nuclei are again graded. It is natural to compare the particular nuclear status introduced in the latter case by a particular kind of radiation (gamma-rays) with a former case of dust-free air in the absence of recognized radiation. In other words, if we abstract from the details of the mechanism which are un- known for the colloidal nuclei, chemical agglomeration might be con- sidered referable to some radiant field, unknown but otherwise essentially alike in kind, to the much coarser nucleations observed on exposure to the known radiant field. The effect of radium, however distant, is virtually 50 VAPOR NUCLEI AND IONS. an increase of size of the efficient air nuclei and a decrease of their number. Hence if we were to fancy that the colloidal nucleation of air responds to its own radiant environment, this would have to be special in kind. Recently I have made similar tentative inquiries as to whether the ions and persistent nuclei might not be regarded as colloidal nuclei aggregated by kinetic pressure, corpuscular or undulatory; for in view of the occurrence of pronounced secondary action within the fog chamber, the radiation at any point must be considered as sufficiently the same in all directions to be equivalent to a Lesage medium. It is then possible to account for the nucleation in any ionized field, for fleeting and per- sistent nuclei, for condensational differences of positive and negative ions, for fleeting nuclei otherwise identical but respectively charged and uncharged, for the destructive effect of low pressure and the counter- action in strong ionized fields, for electrical differences in the effect of ultra-violet light and of X-rays, etc., with a single straightforward hypothesis. While such a view may be worth a statement, it would at the outset encounter very determined opposition; and the distinctive or differen- tiating evidence to sustain it is uncertain. Briefly, if we admit that with ions sufficiently large and sufficiently numerous, relatively speaking, the colloidal nuclei are virtually non-existent (so far as the fog chamber is concerned), since the former capture all the available moisture, most of the phenomena of nucleation admit of interpretation, and additional hypotheses, however alluring, are not called for. Moreover, so long as the representative colloidal nuclei are definitely smaller than the smaller ions, even in the strongest electrical field — in other words, if what may be called the shattering action of strong fields always fails to reveal ionized nuclei as small as the representative colloidal nuclei — the special interpretation is not warranted. In conclusion, if we asked what is the most important outcome of researches of the present character, I should refer to the appearances obtained, indicating that a gas, or at least a moist gas, far from being a uniform system, behaves like an assemblage of nuclei which decrease in size and increase in number as the molecular dimension is approached. Every group of nuclei is none the less a structurally essential part of the gas and (be the number of groups few or many) is at once restored if withdrawn, while the molecule itself is distinguished among the many nuclei of its own kind by the maximum frequency of occurrence. In the above cases (fig. 25) the occurrence of nearly identical slopes for colloidal nuclei and for strong ionization may thus be regarded as the initial branch of a law of distribution of sizes given by the theory of dissociation. EFFICIENCY OF FOG CHAMBER. 51 COLLOIDAL NUCLEI IN DUST-FREE AIR. EXHAUSTION PIPES AND STOPCOCKS FOUR INCHES IN DIAMETER. 44. Purpose. — In the above experiments, the efficiency of the fog chamber in regard to the capture of very small nuclei was successively increased by enlarging the efflux pipe in diameter as far as 2 inches. In the present experiments a further step is taken by increasing the diameter to 4 inches. This final step, however, did not show the marked improvement which might have been anticipated, while the difficulty of manipulating a plug weighing 25 pounds is obvious. The limiting coronas here, as above, were the large green-blue-purple type, and in no apparatus of the present kind has it been certainly possible to exceed this in angular aperture. The use of so large a stopcock introduces other difficulties. It is in the first place nearly impossible to make it quite tight. Provision against the influx of external air may be made in perfection, by aid of an annular oil bath of the above form; but slow leakage from the fog chamber to the vacuum chamber around the plug could not be avoided. This is an inconvenience, though it need not introduce serious error. With a tightly packed filter the air within the fog chamber is never quite at atmospheric pressure, but a few milli- meters below it, so that the exhaustion begins at a lower pressure than 76 cm. 45. Apparatus. — This is shown in fig. 26, where F is the fog chamber, V the vacuum chamber, C the 4-inch stopcock between, P the air pump. The pressure within the vacuum chamber is given by the gage G, which also communicates with the air pump. The latter may be shut off by the glass two-way stopcock c, which serves also for the admission of dust-free atmospheric air when pressure is to be lowered through the filter /. The gage g is in communication with the fog chamber by the rubber pipe ab, which contains a lateral branch (not shown) with a fine screw stopcock through which thoroughly dust-free air may be admitted into the fog chamber from a long well-packed filter (not shown) beyond the cock. The heavy plug of the stopcock is handled at h and counterpoised by the spring p on a pulley. By properly adjusting the tension (taking care to allow for excess and diminution of air-pressure), the plug may be rotated as easily and quickly as a much smaller valve. There is no evidence that increased speed in the rotation of the stopcock would have increased the efficiency of the fog chamber. The limit reached depends rather on the law of flow, the gradient of which eventually vanishes. The projecting rims, m and n, of the stopcock and lower end of the plug, form annular troughs into which oil or mercury may be 52 VAPOR NUCLEI AND IONS. poured (iron stopcock) and the inflow of the external air absolutely avoided. The cock, moreover, is in this way virtually floated in oil; though no such provision is possible to avoid leakage from F to V through C as already stated. The goniometer for the measurement of the angular diameter of the coronas is shown somewhat indistinctly at D. The small vertical plate which serves as the eye rest is 40 cm. from the axis of the fog chamber, to which it is rigidly attached, but with freedom of rotation about a vertical and horizontal axis and translation along the latter. The arms are 30 cm. long, opened and closed by a tangent screw. The point source of light (not shown, part of a Welsbach mantle) lies 250 cm. behind the fog chamber, in the same horizontal plane. Fig. 26. — Disposition of apparatus in case of fog chamber (F) and vacuum chamber (V) connected by a 4-inch exhaustion pipe, etc. The X-ray bulb, adjustably placed with the induction coil and inter- rupter on a table provided with wheels, could be moved as near to the fog chamber, or remote from it, as desirable, with facility. The fog chamber itself was a cylindrical jar of glass, provided with a metallic face held in place by bolts and tightened by a rubber gasket. Wax-resin cement was used in liberal quantity everywhere, and internal metallic parts were so far as possible covered with a coat of it. In later experiments a cylinder of wet cloth was placed in the tube between C and F, to saturate air, in addition to the rectangular framework within the fog chamber. EQUATIONS OF FOG CHAMBER. 53 46. Exhaustion difficulties. — As the stopcock was not tight internally, the final or isothermal pressure in the fog chamber could not be regis- tered. In any case it is doubtful whether the cock can be closed again quickly enough. It was therefore customary in the following experi- ments to put the vacuum chamber and the fog chamber in contact for a short time, after isothermal conditions had been reestablished. Long communication between the fog chamber and the vacuum chamber is unadvisable from other points of view, since the character of the nuclea- tion of the latter is not so well guaranteed It is therefore necessary to ascertain the conditions under which exhaustion takes place. Let v be the volume of the fog chamber, V the volume of the vacuum chamber, k/c the ratio of specific heats of the moist gas, and let p, v, t, p, denote its pressure, volume, absolute temperature, and density, under conditions given by the subscripts. It will be convenient to refer to the vacuum chamber by the same symbols with accents. Hence the successive thermal states will be for dry air, For the fog chamber . For the vacuum chamber . Initially ^=76, » p Adiabatically (alone) p x r, p, Isothermally(alone) p 2 t 2 = t p.,= p t Isothermally (together) . . . p 3 r :i = r p 3 Initially p' t' = t p' Adiabatically (alone) .... p' ', =- p i t' x p\ Isothermally (alone) . . . . p' 2 t' 2 = t p' 2 = p' Isothermally (together) . . p' 3 = p 3 r' 3 = r, p' 3 = p :i The equations describing the transformations are (again for dry air) (k-c)/k [p =Rpr p' =Rp'r I p^Rp^ p'x =Rp\t p 2 = Rp 2 r = R Pi t p'i = Rp' 2 r = Rp\r p 3 = Rp 3 r orp 2 /p' 2 = P2/p'2 3. ■ • .Pv + P 1 V=P 1 v + p'\V = P 2 v + P r 2 V=P 3 (v+V) From these one may deduce, relative to the value of p v v (p(k-c)/kp iC /k- p 3 ) = V (p 3 - p'(k-c)/kpc/k} or C /k- ^(l+W^) 4 Pl ° p'(k-c)/k+ (v/ v) p {k-c)/k 5 p'% + p2'V/V = p t (i+v/V) 54 VAPOR NUCLEI AND IONS. where p, p', p 3 , are observable with certainty. While equation (5) is variously useful in checking the results, it does not admit of the indi- vidual determination of p 2 and p' 2 . For this purpose, however, the equations (1) and the second and third of group (2) are available, with the results (for dry air) 6 p 2 =p(k-c)/kp t c/k p' 2 =p'(k-c)/k pi c/k since p x c / k is given in equation (4). Using these equations, the data of table 17 were computed, in con- nection with incidental results tested for the purpose. Table 17. — Values of adiabatic (/>,) and isothermal (p 2 , p' 2 ) pressure for dry air. Pres- sures computed from the initial pressures (/> = 76, p') and the final common isothermal pressure p 3 of the communicating fog and vacuum chambers with other similar data. Volume ratio v/V =0.064; temperature 20 . Observed. Computed* Observed. r i-'i = 273 t'-* 2 ~273 P- P'- Pz- t#V Pv p 2 . p\. XP"r K ft 76 43 39 42 43 45 47 5i 53 55 57 59 43 43 5 5 5 5 5 5 5 5 5 5 5 5 454 42.2 45-2 45-5 47-3 48.9 52.5 54-3 580 59-7 45-7 45-4 48.0 45-2 47-9 47-9 49-3 50.9 54-3 56.1 59-8 62.2 48.4 48.0 45-7 45-7 52.4 59-4 52 58 63 9 3 8 45-0 52.1 59-4 45-6 47-4 48.9 _20.2° - 9-9° ".2° 24-3° 2^5° 20°" * Checked p' 2 + p 2 v/V = p 3 (i + v/V). t Observed as soon as possible at the fog chamber. Nearly same later. t Observed as soon as possible at the vacuum chamber. v"\ =» Pz nearly. The data of table 17 are computed for dry air throughout and are given in fig. 27a, graphically.* The results of the table are very impor- tant. In the first place, it will be seen not only that isothermal pressures or nearly isothermal pressures are not observed, but that the effect of the vacuum chamber is preponderating. Thus the pressure at the latter (p'\) read off as soon as possible and nominally adiabatic is within 1 mm. of p z . Similarly the computed adiabatic pressure (p t ) is within a few ♦The curve (p 2 ) is of no interest here and should be disregarded. EQUATIONS OF FOG CHAMBER. 55 millimeters of p'\ and p 3 . It follows, therefore, that even an approach to isothermal pressure can not be discerned at the fog chamber at all, to say nothing of adiabatic pressure; or that before the exhaust cock can be closed again, the vacuum chamber has practically regained its isothermal pressure by cooling and that the fog chamber is further exhausted by a corresponding amount. The pressure p\=p" 2 observed under isothermal conditions* at the fog chamber exceeds p x (computed) by about 1.9 cm. on the average, which might be regarded as the average vapor pressure of water at the temperature at which the observation was made. Leaving this for further consideration, the final result of impor- tance is the following: p 2 , the computed isothermal pressure in the closed fog chamber, is from 2 to 5 cm. above the observed (nominally) isothermal pressure p\—p 2 (observed), and correspondingly more than this above the common isothermal value (p 3 ) usually taken. For the region in which colloidal nuclei lie the correction will be in excess of the difference between the pressure regions containing Wilson's data for colloidal nuclei, as reduced elsewhere,! and the region in which my own data as summarized below would lie. In other words, the data in my large coronal apparatus lie in regions of exhaustion at least as moderate as those observed in Wilson's small apparatus; or the two types of apparatus compare in efficiency if the drop of pressure taken is not the apparent experimental value but that deduced for the computed isother- mal pressure (p 2 ) of the fog chamber, as above explained. 47. The same, continued. Case of air in the fog chamber saturated with water vapor. — It will next be necessary to compute the above data with allowance for the vapor pressure of water in the fog chamber, supposing the vacuum chamber to be dry, which may seem to be in a measure true, since it is heated by the same transfer of air which cools the fog chamber. Hence in the summary, if n is the vapor pressure of water, the equations relating to the fog chamber have to be modified to p ^ n \(k-c)/k T, \pi-nj p —Ti =Rpz px~~ Xl = RpiTl P 2 —7T = Rp 2 T = Rp l T p 3 -n =Rp 3 z p 3 = p' 3 = R P ' 3 z 3 pv+p l V = etc.=p 3 v+p\V *This pressure varies but slightly. t Presidential address, Physical Review, xxn, 1906, p. 107, 56 VAPOR NUCLEI AND IONS. From these the value of p i appears as c/k = P*+(P»-*)-v/V p^k-c)/k^( p - n ){k-c)/k^ -njpjc/h . v /v where nj p t occurring with the factor v/V may be neglected Further- more 8 (p 2 - 7z) = (p-7t)( k - c )/ k (p 1 -7: l ) c / k ; p' 2 = p l (k-c)/k Pi c/k where % t (respectively 7r = 0.10, 0.24, 0.51 in comparison with p t = 45.6, 52.2, 59.3) may again be neglected. The results of this computation are given in table 18 and fig. 276. Table 18 also contains the value of dp observed as p-p z and computed as p-p2, the latter being the correct value. This table of corrections will not apply below the limits of computation, since rc l now becomes appreciable in comparison with p x ; but above p-p 3 = 1 5 cm., a table may be constructed from which dp=p-p 2 may be taken at once. In this way the following data will be corrected Since the main data for col- loidal nuclei lie between £-£3 = 25 cm. and 32 cm., the correction will lie between -4 cm. and -6 cm. Table 18, corresponding to table 17 for saturated moist air in the fog chamber. Tem- perature 20 ; vapor pressure 1.7 cm.; £ = 76; v/V =0.064. dp. Observed. Computed. P~p 3 - P-Pr ft'. ft. */>/• t/V'- ft. ft. ft*. Tj. *v 30.5 23-5 16.3 25.2 19.9 14.4 43-5 51-5 59-5 45-5 52.5 59-7 47-9 54-3 62. 2 45-6 45-6 52.2 59-3 50.8 56.1 61.6 44.9 52.0 59-3 -18.7 ~ 8.5 + 1.5 + 24.0 21 . 2 + 20 * Observed as soon as possible at the fog chamber; nearly same later, t Observed as soon as possible at the vacuum chamber; P\"=Pz nearly. Table 19. — The results are for ir = 1 .7 at 20 . p- />'. ft. ft. v/V -i. v/V" =0.064. v/V-O. 76-lr 43-5 59-5 45-5 52.5 59-7 48.1 54-4 60.7 50.8 56.1 61.6 53-i 58.4 63-7 EQUATIONS OF FOG CHAMBER. 57 Fig. 27a. — Computed pressures in the fog chamber and in the exhaustion chamber for atmospheric pressure, p in the former and p' in the latter. Dry air. Table 17. Fig. 276. — Wet air in the fog chamber. Table 18. Fig. 28. — Isothermal pressure in the isolated fog chamber for different volume ratios (v/V) of the fog chamber and vacuum chamber. Table 19. 5» VAPOR NUCLEI AND IONS. It may be of further interest to compute the values of p-p 3 for v/ V = i (a small exhaustion chamber) and for v/V — o (one of immense size) for the same values of p, p', p 3 , observed isothermally and initially at the fog chambers and vacuum chambers and finally when both are in communication. The results for 20 are given in table 19, page 56. These data are charted in figs. 28 and 30, to a first approximation, in terms of v/V and p 3 . If expressed in terms of p', they lie nearly on straight lines, as do the above values of p 2 for v/V = 0.064. The large 45 50 55 60 cm. Fig. 27c. — Wet air in both chambers. Table 20. variation between v/V = and 0.064 as compared with the small one for v/V = 0.064 and 1.0 is noteworthy. Hence, in using small vacuum chambers, the observed and computed p 3 and p 2 lie more closely together, but other disadvantages supervene. The rate of variation 9(v/V) 1 - (p/p') (k-c)/k - {p/p')( k - c )/ k .v/V} 2 is always negative and proportional to p 3 and (p/p')( k ' c )/ k . EQUATIONS OF FOG CHAMBER. 59 10 ZO 30 40 42 44 46 48 50 52 Fig. 29. — Corrections (p 2 - p 3 ) for dp = p — p z . Table 20. Fig. 30. — Isothermal pressure p 2 in the isolated fog chamber for different isothermal pressures p 3 of the communicating fog chambers and for different volume ratios (v/V). Table 19. 48. Case of saturated air in both chambers. — Here all the pressures of both chambers must be reduced by the corresponding pressure of saturated vapor, except p\ where the vapor is slightly superheated. Omitting this the equations become 10 (pi-n) = (p-n) ( * _c) /* (Pi -7: 1 ) c / k or the equations again take the original form, though a special computa- tion is needed, since a different initial pressure (p) enters. These results are given in table 20 and in fig. 29. Table 20. — Corresponding to table 18 when both fog chamber and vacuum chamber are saturated. v/V = 0.064; temperature 20 ; £=76 cm.; ir—1.7 dp. Observed. Computed. P-P+ P~K y. & */>',. Pi* v t/V />'«• Ti- T 2 - 30.5 23-5 16.3 23-7 18.2 12.7 43.5 515 59-5 45-5 52.5 59-7 47-9 54-3 62. 2 43-5 50.3 57-3 52.4 57-8 63-3 45- 1 52.2 59-5 -22.5 -11. 5 - 1-4 * Observed as soon as possible at the fog chamber. t(*,-ir) + -Jr(*-ir)-(i+ £)(*--) 6o VAPOR NUCLEI AND IONS. P-Pz. P~P2. Ratio. o O 16.3 12.7 18.2 23.5 23.5 18.2 ■ 30.5 23.7 11.0 14.2 0.7745 0.7782 P-Pz. Correction. Ratio O 16.3 3.0 23.5 Mean 0.776 23.5 5-3 30.5 6.8 "-** 14.2 0.226 0.225 Mean 0.225 In view of the low temperature (fj) the vapor pressure (7^) may be neglected for the fog chamber; but this would not be the case of the vacuum chamber, where n\ is quite appreciable. As the result of this, the march of pressures in the vacuum chamber is peculiar, but need not be considered here, where p, p', p 3 , and p 2 are chiefly in question. The difference between dp = p - p 3 as observed and dp c = p -p 2 as computed now actually vanishes with the former (see fig. 29). In other words, very nearly ±Z£l- 0.775, or tltl P~P3 P ~P3 0.225 and therefore dp -dp c =0.225 dp , nearly. Hence this is a correction to be taken by preference. A table giving 0.225 dp for the usual ranges of observation was therefore drawn up and used throughout the follow- ing work, or the factor 0.775 may be used at once. These corrections are quite sufficient to indicate that the efficiency of the fog chamber as used above is not surpassed by any apparatus. The preceding correction, in comparison with the cases of sections 46 and 47, seemed to me to be most nearly in keeping with the actual state of the case. The more nearly rigorous solution, when the air in both chambers is continually saturated, leads to transcendental equations for the adiabatic pressures (pi — p\), which can only be obtained approxi- mately. If the vapor pressures {n x and n' x ) correspond to p x and p\, the results would be (Pi-*\) e / k {p x -n x Y/ h = (/>,-*) (1+1//V) (//-7r)(*- c )/* + v/V.(i-(w 1 -w , 1 )/(^ 1 -w , 1 )) c /*(p-7r)(*- c )/* (P S -7Z) (I+V/V) (^An^-n^lip.-n^^Hp' -^ k - c )/ k + vjV .{p~n)(< k - c y k where approximate values must be entered for n 1 \, 7T t , p u in the denom- inator on the right side of the equation. EQUATIONS OF FOG CHAMBER. 6l Similarly p2-x=(Pi- ^i) c/k (P ~ 7r)(*- c )/&, ft - *=(p'i - n\YHp' ~ *)W*, */*-((*-*)/(*-*,))(**»/* r/T/-(^/-3t)/(^/-*/))<^)/* Making use of the values of table 18, the data of table 216 were com- puted on a single approximation. Table 216.— Corresponding to table 18, when both fog chamber and vacuum chamber are saturated. v/V =0.064; temperature 20°; ir = i.7; £=76 cm. *ir. *<• obs. P'- obs. h> p 2 . *V *« f%> OO 2-3 43-5 45-5 46.1 54.7 44-9 -17.8 + 24.1 .2 1-9 51-5 52.5 52.5 59-6 52.0 - 8.3 21.3 •5 1-7 595 59-7 59.3 64.6 59-4 + .8 19.8 ♦Assumed from data for t\ , iT\, in table 18. P-Pz- P~Pr (P-PJAP-P*)- IT. irV h- Pi/ Pi- ?r 0.0 16.3 23-5 30.5 0.0 1 1.4 16.4 21.3 16.4 -2£- .69 14.2 O.7 9 2.2 i-9 1-7 + 5-2 + 9.4 + 12.7 0.910 .929 •952 499 55-5 61.5 The corrections (see table 216, and upper graphs in fig. 30) again lie on a curve which passes through zero, but with a larger slope. In fact, they are so much larger than the preceding cases and throw the whole phenomenon into so low a region of pressures that it has seemed best to abide by the data at the beginning of this paragraph, at least for the present. Details are given in table 216. A few incidental results deserve brief attention. The first of these is the nearly constant difference of about dp 2 = 2 cm. between the ob- served value of p 2 (nominal) and p 3 . Since for dry air or not (p' 2 -7t)+v/V. (p 2 -7:) = (p a -7:)(i+v/V) is constant for a given exhaustion, dp' 2 = -v/V. dp 2 . Hence if dp 2 = 2 cm. dp' 2 = o. 064X2=0. 13 cm. 62 VAPOR NUCLEI AND IONS. The case is illustrated graphically for p' = 45 cm. in the notched curves of fig. 27c in a way easily understood. It seems probable that whereas the smaller fog chamber has more than returned to isothermal conditions (p 2 ), the large vacuum chamber is about a millimeter short of it when the cock is again closed. The constancy of this difference is in all probability referable to the systematic method of investigation, though the effect of precipitated moisture has not yet been considered. Anomalous relations in the data for the fog chamber (as in the case of ^ = 59.5 cm.) are direct errors of observation. On the other hand, however, since within the ranges of observation and very nearly p =a p2= a 2+b 2 p' p 3 =a 3 +b 3 p' (P ~ P%) /(P-Ps) = (A s + B z p') /(A 2 + B 2 p') =A+Bp' (nearly) , where a, b, A, B, etc., are constant. Frequently B is negligible, so that (£-£ 2 )/(£-£ 3 )=^ = const., in which case the graph for (p 2 -pa)/(p-p3) — i-A also passes through the origin as in the two cases (fig. 30). There is no need of this and it is at best an approximation which facilitates computing. Some remarks may here find place on the moisture precipitated in the fog chamber per cubic centimeter, and on the absolute temperature x x to which this precipitation heats the chamber above the adiabatic temperature z v I have shown above that the combination of fog chamber with a large vacuum chamber and a sufficiently wide passage- way, though affording superior practical advantages, and little if any inferiority in efficiency to the piston apparatus within the range of measurable coronas, nevertheless does not give the volume expansion of the air within the fog chamber either under adiabatic or under iso- thermal conditions. It makes no difference how rapidly the stopcock is manually closed. The conditions of the vacuum chamber are always impressed upon the fog chamber. The adiabatic and isothermal data may, however, be computed if the volume ratio of the fog and vacuum chambers and the pressures before exhaustion (when the chambers are isothermally separated) and after exhaustion (when the chambers are isothermally in communication) are known; and these computations are simple because the reductions are practically linear. When the vacuum chamber is large, moreover, its pressures vary but slightly, and therefore the pressure observed at the vacuum chamber after exhaustion, when the two chambers are in communication, is very nearly the adiabatic pressure of the fog chamber. This result makes it EQUATIONS OF FOG CHAMBER. 63 easier to compute the water precipitated per cubic centimeter (without stopping to compute the other pressures) with a degree of accuracy more than sufficient when the other measurements depend on the size of coronas. To prove this, let d, L, and n refer to the density, latent heat of evaporization, and pressure of water (or other) vapor; let p, k, c, t, denote the density, specific heat at constant pressure, specific heat at constant volume and temperature of the air, the water vapor contained being disregarded apart from the occurrence of condensation. Let the variables, if primed, refer to the vacuum chamber, otherwise to the fog chamber. When used without subscripts, let them refer to isothermal conditions or to room temperature. Let the subscript x refer to the adiabatic conditions on exhaustion; the subscript 2 to isothermal con- ditions, if the chambers could be isolated immediately after exhaustion and allowed to heat and cool from the adiabatic state independently. This case is in fact realized in the piston apparatus. Let the subscript 3 , finally, refer to the isothermal conditions which prevail if the cham- bers are put in communication at a room temperature after exhaustion. Then the usual equations for heat evolved in the condensation of vapor lead easily to where d is the density of saturated vapor at t, where t is the tempera- ture to which the wet air is heated from its adiabatic temperature t v in consequence of the condensation of water vapor, where [dj is the density of water vapor if cooled as a gas, i. e., without condensation, from t to t v Moreover [dj-df-m (2) the mass of water precipitated per cubic centimeter by condensation or the quantities sought. If Boyle's law is assumed to hold both for the gaseous water vapor [d t ], and for the wet air, it is convenient to reduce equation (i) to room temperature (isothermal state) and it becomes If the vapor density of saturated water vapor is known at a temperature as low as t, d=f (T) (4) so that t, the only unknown quantity in equation (3), since the equation of adiabatic expansion determines t v is found from the intersection of the curves (3) and (4). This is the method of Wilson and of Thomson. 64 VAPOR NUCLEI AND IONS. In the piston apparatus p 2 as well as p may be read oft' by the gages; but, as stated above, this is not true when the fog chamber communicates directly with the vacuum chamber. In this case, however, p x is nearly given by p 3 . Consequently it is expedient to reduce equation (i) to the adiabatic conditions, whence (if t refers to absolute temperature), <*=<* (^r /k -^(^y /k (i- ti ) (5) Here n v the vapor pressure at t v is usually negligible (about 0.5) as com- pared with p v and p t may, in practice, where great accuracy is not demanded, be replaced by p 3 , which, like p, is read oft while n holds at t, which is also read off. In illustration I will give a numerical example taken from table 2 1 b where p = 76 cm. Pz - 45-5 cm. t x =-17.8° C. ir = 1.7 cm. £1 = 46.1 cm. i =5-25° C. P =0.00118 /> 2 = 54.7cm. *' 1 =2 4 .I°C £'=43.5 cm. If equation (3) is taken ?w=<;.36Xio"~ 6 grams per cubic centimeter If equation (5) is taken w=5-35Xio- 6 If equation (5) is taken and p x replaced by p 3 w*=5-3oXio~ 6 the error being 1 per cent of the true value, which is quite near enough in practice and admits of easy correction. Finally one may compute f 2 , the pressure which would be observed at the fog chamber instead of p 2 if allowance is made for the water pre- cipitated in the fog chamber, whereby additional air escapes into the vacuum chamber, since the former is heated from Tito z x . The chambers are supposed to be separated (cock closed) immediately after conden- sation and no further loss of air is to take place from fog chamber to vacuum chamber while the absolute temperature of the fog chamber passes from ij to atmospheric z. Without giving the full discussion for which there is no room, I will merely add that P2-k = J-(Pi— *) EQUATIONS OF FOG CHAMBER. 65 where 7 } is the density of the air at t,. The ratio is equal to * (ft-^i)(T 1 /T 1 + V^T , 1 /T 1 ) if the vapor pressures 7r are designated like the temperatures and pres- sures with which they are associated. Usually * tl/Xt+v/V.-fJlTi suffices, the term involving the vapor pressures (n) being a correction of about 1 per cent. The computation, to which I shall return elsewhere, shows that p 2 , computed, is always above p 2 , observed, so that the fog chamber begins to heat itself above the temperature v t before the cock can be closed again and contains less than its normal allotment of air. Thus, in the example given, p i /pi = o.gi\ £2 = 47. 9, observed; and p 2 = 49.9, computed. Hence p and p 3 alone have any definite meaning for the fog chamber. 49. Observations with 4-inch exhaust pipes. — The observations of this paragraph, apart from the exhaustion difficulties already discussed, were made peculiarly difficult by the unavoidable leak from fog chamber to vacuum chamber through the large exhaust cock. It was therefore essential to wait many hours for each observation, since the coronas corresponding to the second, third, etc., of successive exhaustions were smaller than the first. This can not be due to any other cause than the presence of water nuclei from the fog of the first exhaustion. A given corona, moreover, was apt to decrease in aperture as much as one-half during the period of subsidence, showing growth of certain particles at the expense of others, the latter being afterwards detected in the water nuclei specified. This must also be attributed to the continued slow exhaustion due to the leak in question. In table 21 and fig. 31 the earlier data with 4-inch pipes are given, chiefly with the object of direct comparison with foregoing results with 2 -inch pipes. The meaning of the data is clear from the earlier tables, and the dp here mentioned is the isothermal value observed at the fog chamber as heretofore. On March 16 to 18 the data are irregular in the way common to observations in a newly adjusted apparatus. The effective nucleation is too small from the presence of interior sources of relatively bulky nuclei. A fairly complete series was undertaken on March 19. The out- going and incoming branches do not quite coincide, and the data, as a whole, still lie below the corresponding results with 2 -inch exhaustion pipes. This is indicated in fig. 31. Of later observations (March 22), 66 VAPOR NUCLEI AND IONS. 25 30 Fig. 31. — Nucleations (w) observed in dust-free air and dust-free X-air at differ- ent exhaustions (dp); 4-inch pipes. Table 21. DISTRIBUTIONS OF NUCLEI. 67 Table 21. — Colloidal nuclei in air. New apparatus. Glass fog chamber. Piping 12 inches long, 4 inches in diameter; 4-inch plug stopcock. I. Mar. 16 Later . . II. Mar. 17 Later . III. Mar. 18 IV. Mar. 19 Incom- ing se- Ob- served dp. 25-4 25.4 25-4 29.0 29.0 20.4 251 25.1 29 29 29 33 33 33 38 38.0 42 42 33-5 33-5 33-5 33-5 33-5 3i-5 30.0 28.2 27.2 26.5 24.1 22.6 20.6 19. 1 17-5 ■7.2 Cor. w y' w y w y wB P wy' w o w y gyo cor w o gyo w pcor w o cor y'obg Stone- blue w y gyo gBP b' B gBP gyo togB gBP wog cor cor cor cor cor cor cor cor nxio- 20 3 20 2 i6 27 o 12 25 5i 20 28 140 3i5 3i5 165 330 35o 350 425 30 3i5 425 190 3i5 4i 390 540 510 500 4 425 460 300 40 18 14 13 12 10 7 2 Ob- served s. Cor. dp. Incom- 16.5 Rain ing se- 151 Rain ries. Later . . . 16.5 Just seen Outgo- 17.8 No ing corona series 19. 1 No after corona several 21 . 1 2.2 cor hours' 23.1 3-6 cor rest. 25.0 3-1 cor 26.5 3-7 cor 28.2 7-i wyg 30.3 9-7 wog' 32.1 11 wybg 32.8 12 gBP nx 10- 3 . Promiscuous results. V. Mar. 20 Later . . . VII. Mar. 22 Later . . X. X-rays ; Z? = 20 cm. 5 29.8 gBP 30.0 yobg 30.3 gBP 28.0 gyo 27 -3 11 wr'g 26.2 wpcor 28.6 •6.9 gBP 26.3 4-5 cor 27.4 5-5 cor 29.9 10 yobg 33-9 11 gBP 33-8 11 gBP 32.7 gyo 33-6 „ gBP 30.6 11 gBP 27.9 11 gBP 25-3 11 gBP 21.7 5 cor 21.6 50 cor 22.5 10.5 wrg' 23-5 gyo 22.8 5-o cor 22.6 10 wyg 21.5 3-6 cor 21.6 3-7 cor o o 3 16 9 19 120 270 390 470 450 37o 450 390 250 165 120 33 62 330 475 475 420 470 455 430 400 39 39 210 340 40 310 16 1 «= 3.8 persistently except on first exhaustion. 2 Many periods follow. 'Blurred; other coronas very clear, often multi- annular. 4 Periods usually omitted. •Fog chamber greasy. • Followed in next exhaustion by small corona. 7 Apparatus cleaned again. 68 VAPOR NUCLEI AND IONS. 400 300 ZOO too Fig. 32. — Nucleations (n) observed in dust-free air and dust-free X-air at different exhaustions (dp); 4-inch pipes. Table 23. DISTRIBUTIONS OF NUCLEI. Table 22. — Systematic work. Third cleaning of apparatus. p = 76 cm. 69 Isothermal dp=*p—p 3 ; dp. s. Cor. n x 10- 3 . dp. s. Cor. n X io-*. I. VII. Mar. 26 24.8 3-4 cor 13 Mar. 30 30- 5 11.5 yobg 330 24.8 3-8 20 30.5 10.7 wog 300 24.8 3-9 20 28.8 6.5 wrg 100 26.6 3-9 cor 21 29.6 10.5 wrg 25 28.7 4.0 23 33-0 13 gBP 470 304 4-7 42 Later . . . 32.2 12 gBP 460 31.2 5-2 w b r 56 VIII. 32.2 8.6 w p 215 Mar. 31 30.8 "•5 gBP 460 II. 30.8 12.5 gyo 410 Later . . . 33-4 13 gyo 425 29.6 ". 5 r y°g 355 33-4 9-5 w p cor 240 27.7 5-2 wr 50 33-4 11. 5 wybg 390 IX. 33-4 9-5 w b r 250 Apr. 1 312 12 gBP 460 304 8-7 wpcor 210 312 13 gyo 410 28.8 50 cor 46 Later . . . 30.6 12 gyo 410 26.8 3-5 cor 15 30.1 n wybg 375 251 2.7 cor 7 X. 22.0 2. 1 Faint 2 Apr. 2 30.7 12 gy bg 410 cor 30.6 12 y'gbg 380 III. 30.6 7.8 w b r b 180 X-rays 21.5 3-8 cor 18 XL from 23-7 11 Vague 345 Apr. 3 29.4 7-7 wPB 147 D=20 gyo 29-5 5-8 cor 74 cm. 22.6 4.4 cor 27 XII. Vague Apr. 4 297 n yobg 300 22.5 6.7 gBP 124 29.6 n wog 300 22.6 8.7 wp cor 175 28.8 8.8 wPB 220 25.5 gFog 400 29.0 2 8-3 wP? 175 IV. 28.0 6.0 cor 80 Mar 27, '30.5 gyo 410 XIII. 2d day. 316 10 yobg 380 Apr. 6 33-9 3 i3 gBP 470 29.1 10 wog 145 26.5 3-9 cor 22 V. 26.9 3-9 cor 22 Later . . . 31.0 12 gyo 410 27.6 5-2 cor 52 30-4 4.8 cor 45 28.8 8-9 wP 215 29.4 7-4 wp 145 27.6 7-3 w y 120 28.7 5-8 cor 72 XIV. 28.0 5-o cor 46 Apr. 7 27-5 7-4 wrg 120 27.1 3-3 12 27 -5 cor 28.0 4.2 27 27-5 5-2 cor 52 28.8 6.0 w y 80 XV. 27.6 5-i cor 49 30.0 7.2 wB P 150 Apr. 8 27.2 2 2.7 cor 7 30.6 8.1 wP cor 180 27.0 3-5 cor 15 3i.4 7.0 w y 130 26.9 4 3-5 cor 15 32.0 11 gy 4i5 25-5 3-o cor 15 33-8 gBP 475 Series VI. Apr. 8 Mar. 28, 34 -o 13 gBP 470 (55) 25- 2 i-3 cor i-5 3d day. 32.0 13 gyobg 4i5 (56) 27.1 2.6 cor 6.4 30.5 12 wybg 380 (57) 28.9 5-9 wrg 78 29.8 8.0 w b r b 180 (570 8.4 wP cor 210 28.8 6-5 wrg no (59) 32.5 12.5 wyg 390 28.0 4.1 cor 24 XVI. 27.0 4.1 cor 24 Apr. 10 Later. . . 305 10.6 yog' 330 (58) 32.0 13 gBP 460 28.7 6-3 cor 88 30.8 13 gyo 455 27.1 3-4 cor 14 XVII. 256 3-4 cor H Apr. 13 30.0 10.5 wrg 255 VII. 30.6 9-5 wrg 240 Mar. 29, 30.3 yobg 330 Apr. 14 307 12.5 ygbg 380 4th clay. 30.6 1 1 yobg 33o 30.6 n-5 wog 300 30.6 9-4 wcg 230 Apr. 15 30.9 13- gytog 455 7° VAPOR NUCLEI AND IONS. while there is evidence of slight improvement, irregularities are at times increased. The final data in the table relate to the effects of X-rays from short distances (D = 2o cm. to anticathode) , and the peculiar feature here is the steepness of the line as compared with earlier results. The reason is in part due to the admission here of dp=p-p 3 instead of the observed dp as heretofore, the effect being to displace upper observations rela- tively more to the left. It is already quite clear, however, that neither for X-air nor for ordinary air have the earlier data been much improved. The highest corona attained in both cases is again the green-blue-purple type. Table 23. — Atmospheric air. Two observations daily. Water nuclei removed by low exhaustion. dp = p- p z . Aug. diameter s/30 nearly. Fog chamber 40 cm. from eye, 250 cm. from light. Date. Observed. Computed. dp. s. Cor. n ) 9j/S J/ Sp* / s 1 I ml '• ' i f 1 H > 71X103 J Jo zo 85 Fig. 33. — Nucleations (n) observed in dust-free air and dust-free X-air at different exhaustions (dp); 4-inch pipes. Table 23. DISTRIBUTIONS OF NUCLEI. 73 52. Discussion. — The new curves (fig. 33) lie nearer together for the energized and non-energized states of the gas. This obviously results from the correction applied to the observed values of dp. Neither do the graphs ascend as highly as they did before, for the same reason, remembering that the nucleation (n) always refers to the exhausted fog chamber. Again, the new curves must be steeper than the old; but both of them, i.e., the curves for the non-energized and for the energized gas, have about the same slope, so far as can be made out for the case of such steep curves as those under consideration. Nevertheless, it is probable that there is some other reason implied in this, which is yet to be made out. For the ionized state the observations frequently suggest a kind of saturation beyond which the ions pass into per- sistent nuclei very much as a vapor condenses. In other words, a maximum ionization pressure, determined by a definite number of ions per cubic centimeter, which can be approached as the radiation is more and more intense, but not exceeded, is a useful conception in con- nection with many of the experiments given. 53. Summary. — The general summary of this chapter has already been given in sections 42, 43, and 52, particularly in the former, with regard to fig. 25, and need not, therefore, be repeated here. The highest order of available coronas has been invaded and surpassed. It appears that the limits of efficiency of the practical fog chamber with rapidly opened plug cock have been reached when the long cylin- drical vessel of about 6,000 c. cm. is exhausted into a vacuum chamber of about 100,000 c. cm. through a pipe not less than 5 cm. in bore nor more than 50 cm. long, with a stopcock of wider diameter (7 to 8 cm.) inter- posed. To test this again, such a fog chamber was adjusted as shown in fig. 34 (V vacuum chamber; F fog chamber; G, g, gages at the former and the latter, the whole mounted on casters to admit of shaking the water in F; goniometer attached to fog chamber). The results are given, both for fog chambers Nos. I and II, in tables 24 and 25, with correc- tions for the barometer as explained in Chapter VI, sections 100 and 101 , and they are charted on a small scale with other data (table 23) in fi g- 35- The graph for chamber number II would coincide with data inferred from Wilson's colors for small coronas; but for large coronas it actually lies in a region of lower exhaustion — to which however, too much importance must not be attached, because of the difficulty of identifi- cation. The point is that the apparatus of 2-inch pipes is quite the equal, if not the superior, of the apparatus with 4-inch pipes, in the region of both the lower and higher coronas. Curiously enough, the apparatus I 74 VAPOR NUCLEI AND IONS. (4-inch cock) gives an excess of ions, whereby the graphs for I and II in the region of lower coronas are consistently distinct. I will pass this over here, as it needs additional investigation. That the limit of efficiency has been reached for plug-cock apparatus is specifically proved by the fact that the same large green-blue-purple Fig. 34. — Disposition of apparatus in case of fog chamber (F) and vacuum chamber (V ), connected by 2-inch pipes and 2^-inch stopcock. corona terminates the observations, no matter whether the nuclei are relatively large, as in case of the ions and intense X -radiation, or relatively small, as in case of the colloidal nuclei. This is differently proved in the work with alcohol, in Chapter IV. It is somewhat hard to understand why the efficiency should terminate abruptly, with a certain number of nuclei per cubic centimeter, no matter DISTRIBUTIONS OF NUCLEI 75 whether large or small nuclei are in question. One should expect these conditions to depend on the size of nuclei ; but (as the data show) even though larger numbers of nuclei are certainly present, they are devoid of efficiency beyond the limit. of 6 \lo.R. Tab. 26,1. " ». » " ,2. " I. " 24. Mo.I. Tab. 26, 3. " " . » " ,4. "omputed. / f \ 1 . c / DUSl -FREE AIR B A / FIG. 3 5 TIX/O' 5 =p-p 2 l_£ '£ [•/' 26 Fig. 35. — Nucleation (», in hundred thousands of nuclei per c. em.) observed in dust- free air and in energized air at different exhaustions. dp = p-p 2 . 2-inch and 4-inch pipes and perfected apparatus I and II. Tables 23, 24, and 25. Ions in No. I con- sistently in excess. Table 24. — Fog chamber and vacuum chamber as in fig. 34, joined by 2-inch pipes, 2^-inch plug stopcock. dp= p - p 3 . Final series for comparison. dp. 5". Corona. P~P2- ft X io -3 . .y cor- rected.* n Xio~ s corrected. I. Barometer 25-9 3-2 cor 20. 1 9-5 3-4 n. 6 76.2 26.9 6.4 wp'g 20.8 76 6.6 83 27.4 6.8 gBP 21 . 2 120 28.7 10. 2 wrg' 22.2 210 29.4 12 yr 22.8 310 30- 5 13? gBP 23.6 380 33-5 13? gBP 26.O 410 II. Barometer 24.9 i-7 cor 19-3 1.8 1.3 1 4 75-5 25.6 3-6 cor 19.8 15 3-2 9 9 26.4 5-6 cor 20.5 53 5-2 42 Table 2= . — The same; 4-inch stopcock, apparatus I. dp. j. Corona. P-Pv nx 10- 8 . s cor- rected.* n X io~ s corrected. III. Barometer £ = 75-8, July 14. IV. /> = 75-9, June 3. 25 -5 27. 2 25.0 28.1 27-3 27. 1 25-9 26.1 27.4 ?3 4 2 7 5 4 2 2 5 7 9 4 2 3 2 8 7 cor y' b P cor 19.7 21 . I I9.4 21.8 21 . I 21 .O 20. I 20. 2 21-3 15 37 4 no 45 22 6 6 40 3-6 4.8 2-3 5-2 * Correction for barometer made as explained in Chapter VI. 76 VAPOR NUCLEI AND IONS. In no form of the fog chamber have the initial yellows and browns of the steam jet been approached. Coronas reach but to the equivalent of the opaque zone. Below the large green-blue-purple corona, the com- puted value of the water precipitated per cubic centimeter may now be considered trustworthy. In Chapter I this was true below the middle green-blue-purple corona. The drop of pressure (p - p 2 ) actually efficient in producing low temperatures in the fog chamber must be computed from the initial pressure of the isolated fog and vacuum chambers before exhaustion, and their final pressure when in communication after exhaus- tion, all data taken at the same temperature. As the intensity of radiation increases, coarser nuclei become much more frequent, but beyond this the coronal method is not adapted to test the lower limit in question. At the upper limit, however, most of the observations show that the finer nuclei become more abundant when the medium is more powerfully energized ; or that new gradations of nuclei of continually increasing smallness and continually increasing number are produced by radiation increasing in strength indefinitely. These are important questions, however, upon which I hope in the near future to make some final tests. CHAPTER III. MISCELLANEOUS EXPERIMENTS. 54. Objects.— Having perfected the coronal fog chamber to the degree specified in the earlier sections of Chapter II, it seemed expedient to make use of it for a variety of purposes partly corroborating and interpreting my earlier work, partly introducing new results. In particular, the growth of persistent nuclei in a highly ionized medium in the lapse of seconds, the occurrence of solutional nuclei, and like questions may be studied by the depression of the terminal asymptote produced by the introduction of relatively small numbers of larger nuclei into the medium. Again, the effects of radiation from different distances on the medium of the fog chamber, the absorption of such radiations, the distributions within the fog chamber, etc., may be elucidated by this treatment. Finally, some consideration of the rates of generation and decay of ions is in place and a method will be shown for the standardization of coronas. Some final remarks will be made on the steam jet and on the relation of its color phenomena to those of the fog chamber. Throughout this chapter dp refers to the drop of pressure observed at the isolated fog chamber under isothermal conditions, the exhaust cock being closed as soon as possible after the expansion. The necessary reductions (should they be needed) may be made as shown in the pre- ceding section. 55. Growth of persistent nuclei. — In table 26 (illustrated by fig. 36) the time during which the fog chamber was exposed to the X-rays, with the anticathode at a distance of D = io cm. from the fog chamber, is given in the first column. The coronas and the number (n) of nuclei per cubic centimeter follow. The two series of experiments made show that there is a gradual increase of the number of persistent nuclei, evidenced by the gradual reduction of the number of efficient nuclei. In less than two minutes, however, the phenomenon becomes more stationary, indicating that the full number of persistent nuclei is being approached, or that there are now about as many made as are unmade per second. It is difficult to follow the phenomenon beyond this, for the coronas now become cam- panulate or otherwise distorted, appearing in association with heavy fogs. The true asymptote is probably far off. These data furnished good illustrations bearing on the remarks of section 42, Chapter II. 77 7» VAPOR NUCLEI AND IONS. Table 26. — Persistent nuclei produced after different lapses of time. Method, depres- sion of asymptote, dp =26.4 cm.; D=»io cm. above side of cylindrical glass fog chamber; cock i§ inches; exhaustion during exposure. Exposure. s. Corona. n xio -3 . sec. 13 gBP 410 30 12 wyg' 340 60 II w g' 270 1 90 L ,7-5 O 140 §1 I 20 117-6 140 If i°o .. 13 i m 1 1 1 wyg' 'wrg 340 230 } 90 mi9t 0) 200 $f 14 £ gBP 410 1 Tempestuous fog and rain, leaving smaller corona. Second exhaustion made for safety, showing small corona. 4O0** 3 *n? FIG. 36 s Ss, nxW 3 t 1 0/ stor { 'on L Exposi re (sec.) 80 Fig. 36. — Depression of efficient nucleation (w) of dust-free air ionized by strong X-rays at a given exhaustion, by accumulation of persistent nuclei in the lapse of time- Table 26. 56. Water nuclei produced by evaporation. — A beautiful method of demonstrating the production of water nuclei in connection with the condensation of fog consists in leaving the cock for influx of air from the filter slightly open. In such a case the fog begins to evaporate as soon as produced, and there will be less loss from subsidence of fog particles, in proportion as the evaporation is more rapid; in other words, as the stopcock is more widely opened. Table 27 shows results of this kind and they are reproduced in fig. 37, where the abscissas are dis- tributive. In every case the efficient nucleation (n) of dust-free air, after complete subsidence of the preceding fog, is much in excess of n f , the nucleation observed when the fog is dispelled by evaporation. The table also proves that the degree to which the filter cock (fine screw valve) is open does not influence the result. WATER NUCLEI. 79 Table 27. — Enlargement of nuclei of dust-free air. Glass fog chamber, single wet cloth partition. op= 30.9 cm.; s found after complete subsidence of preceding corona ; s' after partial subsidence with evaporation. Date. s. *'. Cock open. n x io~ 3 . n' x 10- 3 . s. n x io~ s . Remarks. f Full subsi- Oct. 17 » 7 .o 3 7-o 7.0 5^8 45 45 45 45 135 135 135 76 2 7-4 7-4 7-5 140 140 140 dence but with quick opening of cock (90 ) thereafter. f Same, but 6.0 45 84 7-4 140 with slow influx. Oct. 19 7-i 45 135 7-3 45 150 7.3 5 5 7 45 45 150 72 4 74 5 6 2 45 45 150 88 4 7-5 6 3 90 90 150 10 2 7-2 90 140 1 After many days' waiting. l w o g. s After few minutes' waiting. sidence. 8 '9.3 wcg 190 Dust-free air not e nergized; dp =31; repetitions without waiting Dec. 13; internal evolution of nuclei has ceased. 115 wyg 335 "•5 wyg 335 Dust -free , air not ene rgized; dp =36 cm. 13-5 g| P 480 135 g|P 480 Evap. nuclei precipi tated at dp =23.2 cm. and then tes ted at dp =31. 1 30 cor 8 2 0.0 3 2 gyo 410 No difference between 7 and 8. 2 Corona enlarged by the preceding precipitation. Some interesting questions present themselves in connection with this work. Are the nuclei holding positive ions different from those 82 VAPOR NUCLEI AND IONS. holding negative ions? Do they retain their charges, or some equiva- lent of the charges? As there is less mobility and slower recombina- tion in cases of ions entrapped by water nuclei, one would infer greater opportunity for the gravitational separation of the equivalents of the positive and negative charges; for it seems improbable that the water nuclei resulting can be of the same size. Finally, in table 31 data have been gathered showing the gradual self- purification of the fog chamber, after cleansing and sealing. Nuclei arising from some internal source cease to appear in the mere lapse of time and without further interference. The coronas s 1 and s 2 are observed in successive exhaustions at the time (days and hours) given. Initially the first corona is smaller in marked degree, owing to the spurious nucleation within. After more than three days, however, both coronas s t and s 2 are identical. Table 31. — Purification of the fog chamber in the lapse of time. dp = 31, cock 2\ inches; piping 12 inches long, 2 inches in diameter. Exhaust Date. Hour. t x . Cor. s 2 . Cor. n, xio~ 3 . n 2 xio- 3 . Nov. 30 Dec. 1 Dec. 2 13 10 a. m. 5 P m. 10 a. m. 6 p. m. 4.6 7-5 8.0 *9-7 9.2 11 -5 2 i4 gBP w P cor wcg' wrg wyg g' 11 .5 12.0 11. 7 11. 4 11. 5 14 wyg' wyg' gyo w g wyg g i P 40 150 180 275 230 305 460 305 380 380 410 305 335 460 1 Vacuum maintained some hours. 1 dp= 36. Internal source of nucleation absent. The source of this transient internal nucleation is difficult to detect. There were no leaks. Oil evaporation if harmful would be continuous. The same differences of fresh and stagnant air in relation to s t and s 2 are always reproduced. There is no evidence that anything in the vacuum chamber is capable of diffusing into the fog chamber. Some agency therefore, which is productive of relatively large nuclei, and the nature of which is not clear, survives many consecutive precipitations. 57. Distance effects. X=rays. — In the following experiments I endeav- ored to overhaul the curious results obtained when the X-rays strike the fog chamber from different distances, without, however, reaching very satisfactory conclusions. The object of the following work is to bring to bear the newly improved means on the problem. In table 33 data are given as obtained with the glass fog chamber, the drop of DISTANCE EFFECTS. 83 pressure (dp) being below the coronal fog limit of dust-free air. Hence there are no complications from colloidal nuclei. In parts I and II of the table the X-ray bulb was not inclosed in the windowed lead vessel specified, whereas in part III this is the case; while in part IV a tin plate has been placed over the window. Table 32. — Distance effects due to X-rays. Fog chamber of glass with 2.5-inch gas cock and 2-inch piping 12 inches long. X-ray, 7 cells. = 23.8 cm. I. Bulb not inclosed; meters D. 0.5 6.0 10. o 6.3 6.2 7-3 74 11 .0 Corona, n Xio -3 . wcg cor w ; B P w' B P gyo 220 so 75 115 120 260 = 20.3 cm. II. Same; bulb not inclosed 5 w B P 7 2 w B P 2 5 5 2 2 6 S 5 5 4 3 4 6 5 ;;;;;; 1 6 4 wr g 6 3 wrg 5 7 2 wyg 7 3 wyg 90 100 42 42 13 «3 15 28 75 70 100 I OS 22.7 ; X-ray bulb in lead case 1 with window 7.5 cm. in diameter. III. Window open Lid off. 600 4.1 cor 400 4.4 cor 200 6.5 wrg IOO 7-5 gBP + 7-5 gBP + 7-5 gBP + 22 27 85 135 135 135 Same. Tin plate 1 over window. IV. Tin window 600 200 100 2.7 4-9 6.9 cor cor wrg 6-5 40 99 1 Lead sheet 0.12 cm. thick; tin plate 0.03 cm. thick. 84 VAPOR NUCLEI AND IONS. ZOO 180 !40 JZO 100 80 60 40 20 \ X \ FIG. 40 \ \ \ ► ^V x, >e>. ^ *»* ^ TtxW 3 ^v^ *g* ?£es,/' L Vfmett rs) O I Z 3 4 5 Fig. 39. — Nucleation (ions) produced by a radiating X-ray bulb (cased in lead or not as stated), at different distances (D) from the glass fog chamber. Table 32. Fig. 40. — The same, illustrating table 15, Chapter II. The curves I and II, fig. 39, for the free bulbs show different nuclea- tions, compatibly with the different values of dp applied, but are other- wise alike in character. They in no way suggest the law of inverse squares. The curves III and IV for the inclosed bulb are again similar in character, but beyond this very little can be stated. If in case of series II and III, for instance, the law * (A+D) 2 = const. is assumed, the constant A would have to be from 1.2 to 2.7 meters in the first case, and 1.5 to 2.1 meters in the second, which is entirely out of the question for a fog chamber less than 0.5 meter long. If the law n G4+D)= const. is used, in the first case (II) A =0.3 to 0.1 meter; in case III, A = -0.2 to - 0.6 meter. This decrease would then be too fast in the first case and too slow in the second, or the decrement in case of the open bulb is DISTANCE EFFECTS. 85 slower and in case of the inclosed bulb (lead box with window) faster than the first power of distance. In fig. 40 I have inserted data incidentally obtained in Chapter II, table 1 5 (upper curve) , together with data of the same kind in the earlier reports (lower curve). The former are the steepest curves obtained and the latter the least so. Here, as elsewhere, it is difficult to conjecture a reason for this apparently erratic difference of behavior, unless it be referred to the facility with which secondary radiation is evoked, and to the degree in which the fog chamber is pervious to it or generates it. Since dn/dt = a-bn 2 = o, where a is the number of ions produced per second, a must vary as the square of the number of ions, n, observed. 58. The same, continued. Small wood fog chamber. — The data of the last paragraph with the glass fog chamber suggest a comparison with the wood fog chamber. The latter is much the more pervious and in the earlier work showed a much smaller distance effect. Table 33 con- tains eight series of results. In series I, fig. 41, the march is not unlike the case for the glass chamber; but in series II the insignificant differ- ence between a bulb distance of 50 and 100 cm. from the fog chamber (curves III and IV) are similar to I, and often betray the incidental weakening of the X-ray bulb. In the series VI to VIII, change of the drop of pressure (dp) was introduced, but the inherent difficulty of coping with the bulb variations is seen in the details in fig. 43. It is probable that the ordinates of the curves V to VIII (figs. 42 and 43) are proportional to each other; but a discussion is beyond my present purpose. The slow order of change with distance should, however, be noticed. 59. The same, continued. Large wood fog chamber. — It is with this apparatus that the coronas of almost the same aperture were obtained in the earlier work, while the X-ray bulb was moved from 1 to 6 meters from the fog chamber. Table 34 and fig. 44, however, show that this result must have been due to other conditions, for there are changes of nucleation here registered amounting in case of the distance specified (1 to 6 meters) to n' '111 = 145/104 at dp = 22\ 77/41 at dp = ig. For greater pressure differences, dp = 25 and 29, the occurrence of terminal coronas would interfere with the comparisons. At higher exhaustions still, the efficiency of air nuclei would be gradually restored, so that the observed nucleation may be greater with the bulb at 6 meters than at 1 meter from the fog chamber (see Chapter II, figs. 24 and '2 5). At "the very low exhaustion dp = 17 the coronas are too small to be serviceable for comparison. 86 VAPOR NUCLEI AND IONS. ZOO ISO 140 120 100 80 60 u X ■ FIG. 41 \ \ j N > ■ s R ■ h \ ^r \ ^•* H TlXlO~ 3 ■r-fe L D in cm \ Air, 4(70 32> / 7 C/77 J. l_ , > ^^£»5 ^^; /00 ?O0 3<7<7 400 500 600 60 ?0 100 ZOO 300 400 500 600 \ \ FIG. 43 T2Z7 \ ThxlO' 3 *\ ^J L 9 A) c/77. Y 100 ZOO 300 400 500 Figs. 41, 42, and 43. — Nucleations (ions) produced by a radiating X-ray bulb at different distances (D) from the small wood fog chamber. Table 34. With regard to the work of table 34 it should be stated that it is always customary to make a second exhaustion to remove the water nuclei left by the first. Again but two distances (1 and 6 meters) were selected to guard against losses of efficiency of the X-ray bulb, so far as possible. The exhaustion was made during the exposure, which was always brief. At the distance D = 6 meters from the bulb, there is a terminal corona at dp — 22 cm.; at D = i meter it appears a little later at 0/7 = 25. (See Chapter II, figs. 24 and 25.) In fig. 45 the correspond- ing points are joined by straight lines, for convenience. 60. The same, continued. Discussion. — Experiments of the present kind are hampered by two annoying difficulties, the first being the variability of the X-ray bulb, the other the tendency of the wooden fog chambers to develop slight leaks which often pass unobserved. True, the chamber is tested by a second exhaustion after each of the coronas measured; but the unfiltered atmosphere entering anywhere is liable to DISTANCE EFFECTS. 87 produce a disproportionate amount of distortion, because of the relatively large size of the nuclei contained. Hence it is of little value to attempt to systematize the above results in the absence of a well-digested Table 33. — Revision of distance effect. X 45 cm. X-rays. Small wooden fog chamber dp =21.7 cm. 5Xi X-rays off X-rays on II. X-raysoff X-rays on III. X-rays on Bulb weaker IV. X-rays on V. X-rays on VI. X-raysoff X-rays on D. cm. 2 ioo 600 600 600 400 400 200 200 100 100 100 100 50 50 100 100 100 400 400 600 600 600 400 400 200 400 600 600 600 400 200 100 50 25 »5 2,S 2.8 9.2 7-i 7.2 3 5-6 6.2 6.1 6-3 6.2 6.4 7-4 7-9 7-9 8-5 4 .o 9.6 9.0 Corona wrg w y' w r g'B g'BP w r g wrg wrg wrg wrg w r w y w y cor g'BP gBP g'BP wrg w b r w r w p w p wp'g wrg wp n xio- 8 7 180 47 55 58 67 65 100 9i 112 118 112 7 180 180 180 180 180 180 103 106 55 7i 68 74 7i 120 85 26 53 61 77 112 144 144 156 o TOO 175 D. 5 Corona. VI. cm. X-rays on 25 9.0 wp 100 8 wB P 100 7 8 wB P 300 7 3 w r g 300 7 3 wr g 600 5 4 wrg 600 5 2 nXio - 175 120 120 100 100 50 45 High dp = 27. 5. VII X-raysoff X-rays on VIII 5 4-6 600 5 8 400 6 7 w y' 200 9 5 wrg 100 11 wyog' 50 11 5 wyg 10 gyo? 10 5 9 w y' 7 6 wr 100 7 2 wr 10 7 4 g'BP IOO 7 5 wyg' 600 8 4 1 cor \ 7 cor •5 cor 400 8 5 cor 200 7 wrg IOO 7 6 wyg 20 7 6 wyg IOO 7 3 wrg 200 5 9 wrg 300 5 6 cor 400 tt 5 5 500 °5 1 600 8 4 2 600 6 4 3 IOO 8 6 4 wrg 6. 70 105 225 280 350 420 59 93 87 109 84 18 3i 35 35 93 98 98 82 59 50 48 38 21 23 76 '=25.3, »=5-7. *0P= 25.3. »= 5-7. *Freshbulb. 3 Battery current weak. 4 Second exhaustion without rays made (after the first) to remove water nuclei. •#£=28.1, «=48, n=s7,ooo; dP= 29.7, »= 5.2, n= 55,000; §p=* 31.9, #=5.6, n= 71,000. 6 Coronas large on the near end and small on the far end of fog chamber. Subsidence oblique. VAPOR NUCLEI AND IONS. Table 34. — Revision of distance effects. Large cm. Exhaustion during wood fog chamber exposure. . 20X12X55 dp = 19.0 cm. ty-17.1. D. s. Corona. n xicr s . D. s. Corona. n X 10- 3 . X-rays off X-rays on IX. 600 600 100 600 600 600 1 .2 5-3 4-5 6.6 5-2 4.0 just seen 1.2 43 26 77 4i 39 18 XII. 600 100 1 .0 2.8 1 .0 5-3 dp = 24.8* XIII. 100 Mr 100 M- 100 600 100 9-5 27.6 10. 29.4 ^0 6.2 10 wrog' wrog wrog wrg wrog 236 254 265 no 265 dp = 21.8 cm. X. 600 100 100 600 600 5-6 8.7 8.8 6.6 6-5 w p w p w r w y 56 145 145 103 103 dp = 25.1 cm. XI. 600 100 600 6.5 9-5 7.0 wog wrog' w y 112 236 119 ♦Terminal corona; same for 7 and 9 cells. theory of the phenomena; but the equations n = n /(A +D) 2 and n = n /(A+D), where A is constant and D the distance between bulb and fog chamber, may be adduced to accentuate the order of values observed. This has been done in case of table 32, showing that for the non-incased bulb even the inverse first power of D varies more rapidly than the observed phenomenon. To a much greater extent is this true for the wood fog chambers. The phenomenon itself is clearly a case of super- position of primary and secondary radiation. The latter, moreover, is furnished not only by the environment of the X-ray bulb, as shown in fig- 39 by surrounding the bulb with a windowed lead case, but also by the immediate environment of the fog chamber. The small distance variation encountered would then seem to be explained by supposing that relatively much secondary radiation is released by relatively weak primary radiation, as compared with a case of strong primary radiation. Under all circumstances the total effect is an integral to be extended over the whole surface and possibly the interior of the room. Finally, one should recall that the rate at which ions are produced by the radi- ation must vary as the square of the number observed. DISTANCE EFFECTS. 89 240 220 200 180 140 120 O (00 200 300 400 500 600 Fig. 44. — Nucleations (ions) produced by a radi- ating X-ray bulb at different distances (D) from the large wood fog chamber. Table 34. 61 . Distance effect and absorption. Radium. — A few experiments were made incidentally with impure radium (10 mg. 1 0,000 X, sealed in aluminum), using the method of depression of the terminal asymptote. In table 36, for instance, results are recorded on the absorption of the y-rays in lead. Fig. 45 shows the results in relation to the nucleation of dust-free air, the efficient nucleation of which is more and more reduced as the intensity of the radiation increases with the diminishing thickness of the absorbing lead envelope. After nearly 1.6 cm. of lead have been penetrated, the distance of the curve from the asymptotic air line is still marked. Results of this kind should furnish valuable data for testing any theory on the distribution of precipitated moisture on graded nuclei under any definite conditions. Table 36 gives an incidental series of distance effects worked out by the same methods. As exhibited in fig. 46, the first series reaches the 9o VAPOR NUCLEI AND IONS. too cms. of FIG. 45 Oust -fr ?e air .8 Fig. 45. — Depression of efficient nucleation (n) of dust-free air, ionized by gamma-rays of radium penetrating through different thicknesses of lead. Table 35. Table 35. — Absorption of lead. 7-rays of radium 10,000 X. D = ioo cm.; dp =31 cm. high asymptote for dust-free air practically at a distance of D = 150 cm. In the second series the asymptote is lower, due to details in the ad- justment in the apparatus, and reached later, i. e. t at a distance of D = 180 cm. between the radium and the fog chamber. This curve has been worked out completely for D = o cm., and shows the very interesting feature of a well- developed minimum. In other words, as the radium is removed from the fog chamber, the ions which at first predominate and capture all of the moisture decrease more and more in number, until the conditions are ripe for the simultaneous condensation of moisture on the colloidal nuclei of dust-free air. In proportion as the radium is further removed, the latter predominates, fully so when the asymptote is reached. The corresponding results when the radium tube is inclosed in a thick lead pipe (walls 5 mm., length 60 cm.) shows a much sharper minimum, occurring at lower exhaustions. Such irregularities as are apparent here may be referred to the unequal distribution of radiation within the fog chamber discussed in Chapter I, sections 4 et seq. All of these curves have a similar bearing on the question of the distribution of the precipitate of graded nuclei. Thickness n Xio- 3 . t s. 00 7.2 131 i-54 6.0 85 6-5 103 6-5 103 0.40 6.0 85 5-9 82 .0 4.9 49 4.8 46 00 7.0 128 DISTANCE EFFECTS. 91 Table 36. — Distance effect and absorption of radium rays. D measured from side. dp = 31. November 2. Lead pipe 0.5 cm. thick, 60 cm. long D. s. wXio" 5 . D. s. n X io~ s . II cm. 180 150 100 50 25 10 2.2 350 300 250 6.2 5-9 4-7 3-7 3-7 4-7 4.8 5.6 6.7 6.8 6.2 90 82 42 21 21 42 46 70 in 116 90 III. In lead pipe . . On top At 00 cm. 145 100 50 25 10 1 5.6 4-7 4-3 4-i 4.0 49 7.2 70 42 32 27 24 49 135 Above fog chamber . . . On glass Radium at 00 120 100 FIG. 46 1/ — n 60 40 20 / '^m L / -^ V / \\ s* ^-O— & ^ nxio'3 IT D in en 5. 60 too leo zoo Fig. 46. — Minima of efficient nucleations observed at high exhaustions with radium at different distances (D) from the fog chamber. Table 36. 62. Falling to pieces of ions in the lapse of time. — Miss L. B. Joslin contributes the following interesting observations obtained by the method of depression of the terminal asymptote. The data summarized in table 37 and fig. 47 were obtained by acting, in the manner stated, on the dust-free moist air contained within a glass fog chamber, with a sample of weak radium (io,oooX, 10 mg.), sealed in an aluminum tube. This was placed on the outside of the chamber in contact with its walls (0.2 to 0.3 cm. thick), and was then removed 9 2 VAPOR NUCLEI AND IONS. suddenly at given intervals before exhaustion. Only very penetrating primary rays (/? and y) are therefore in question. The curves show the number of efficient nuclei in thousands per cubic centimeter, observed after the lapses of time shown by the abscissas, and it is supposed that the nuclei are reproduced faster than they can be removed by the ex- haustion. In the upper curve the pressure differences applied (^ = 31) are much above the fog limit of dust-free air, which is below d£ = 24 for the given apparatus. In the lower curve the pressure differences are nearly at the fog limit of dust-free air, while the other curve (dp = 28) applies for intermediate conditions. The effect of the radiation is there- fore, virtually at least, a coagulation (to use a figurative expression) of Table 37. — Falling to pieces of ions produced by radium. Everything ready; cocks closed before removing radium. Date. Oct. /.* Cor. MX IO" h-31 sec. CO 7-4 w' obg O 5 8 cor 15 4 O cor 30 5 cor 60 5 9 coyg 90 6 7 wog 120 7 2 wog 180 7 5 wog ^ = 24.5 cm. Oct. 24 6-3 5 5 1 wrg 10 4 7 wrg 15 3 7 20 3 6 25 3 2 30 2 3 5 4 8 wog 10 4 2 wrg 12.5 3 9 wrg 140 76 24 51 82 III 134 145 79 45 36 18 16 11 3- 45 29 23 Date. /.* Cor. MX IO" = 28 cm. Oct. 25 sec. o 15 30 45 60 120 180 00 00 5-8 3-8 3-5 3-3 34 4-7 4.6 4.6 4.8 wrg wog 72 21 16 13 14 40 37 37 43 Oct . 2 7 Repeated . Decay of ions at dp — 24 Radium on . . . Radium at 00 . Radium on . . . 60 2 i-3 30 3-3 t3Q 3-2 15 4.2 10 4-7 6.2 2.5 1.6 12. 1 11. 4 25 36 77 Fog limit of air. Radium off. dp 23.6 22.6 * Seconds elapsed after removal of radium. t First coronas dense but fall out rapidly, leaving fainter and smaller coronas behind. DECAY OP IONS. 93 the colloidal nuclei of dust-free air into the aggregates much larger in size representing the ions. Hence in the presence of radium under the given conditions the number of efficient nuclei decreases either because the ions from their size capture all the available moisture more and more fully, or because the colloidal nuclei have actually been aggregated into fewer but larger systems, which will in turn fall apart in the absence of radium. 120 100 zoo Fig. 47. — Efficient nucleation observed within the fog chamber at different times after exposure to radium applied outside at different exhaustions (dp). Table 37. It follows from what has been stated that above the fog limit of dust- free air the number of efficient nuclei must increase with the removal of radium at a rate which corresponds to the falling to pieces of the ions. The peculiar feature of the results here in question is the manner in which the efficient nucleation decays from the coarser ionized to the finer non-ionized colloidal stages, when the pressure difference is decidedly above the fog limit of air, so that the latter may be recognized. The curves invariably pass through a minimum when the time after the removal of the radium, i.e., the interval of decay, increases indefinitely. This minimum, moreover, is very sharp, almost cusp-like, as if one law were passing abruptly into another. Thus below the minimum (/ = i3 sec, about) the curve for dp = 31 nearly coincides with the curve for dp = 24, which is practically independent of the colloidal nuclei of air. The decay may be computed to be of the order of that of ions. After a lapse of 13 seconds the effect of colloidal nuclei is marked for dp = 31; and even after a lapse of 60 seconds, when the ions (lower curve) have 94 VAPOR NUCLEI AND IONS. vanished to a few hundred, the upper curve is only half way on its march toward the asymptote. This shows the remarkable sensitiveness of the method as a test for the presence of ions or of any nuclei larger than the colloidal sizes. Moreover, measurement of the large coronas is relatively easy. Finally, the curve dp = 31, if prolonged backwards, would seem to start nearly from the origin ; in such a case one would have to picture to oneself a single particle breaking to pieces, in the absence of radiation, into fragments of continually decreasing size, until the debris ultimately numbers 150,000 colloidal nuclei. The intermediate curve (dp = 28) also coalesces* approximately with the other curves for lapses of time less than t = 1 3 seconds. It has its own minimum, however, and from the lower pressure difference, neces- sarily its own asymptote at n — 40,000, since only the coarser order of air nuclei fall within the given limits of condensation in the apparatus used. For the same reason the minimum is lower and later, seeing that the ions are present throughout in relatively greater numbers as com- pared with the efficient colloidal nuclei, than was the case at dp = 31. The curves as a whole have so close a resemblance to the data inves- tigated in section 61 for the effect of radium at different distances from the fog chamber that the same cause must underlie both series of observations. In the former case (distance effects) any given intensity of ionization between the maximum and the vanishing values may be maintained indefinitely by properly placing the radium tube; in the latter case (decay) all stages are passed through in 2 or 3 minutes. Beginning with dust-free non-energized air, the number of efficient nuclei decreases as the number of ions increases (for either or possibly both of the reasons already given) until the condensation takes place wholly on ions. For greater intensities of ionization the number of ions must increase further, and hence the efficient nucleation rises again while the curve passes through a minimum. I , The curves enable us to make certain interesting comparisons, inas- much as the same nucleation results from radium decaying for a stated length of time, as results from the action of radium at a certain distance from the line of sight. From the importance of secondary radiation in connection with these observations, such comparisons are probably not simple. The essential feature is the passage of the nucleation through the same stages of variation, whether of size or of number, in both cases no matter how the given successive intensities of ionization may be produced, or whether they come from within or without. * Considered relatively to the wide divergence after ^ = 13 sec. is passed. The coales- cence need not be perfect. Small coronas fall out too rapidly for close measurement. DECAY OF IONS. 95 63. Decay curve. — Assuming that the rate of decay in the lapse of time (t) is as the square of the number, or that i/n — i/w' = 6 (/-/') where 6 is constant, a few incidental attempts were made to compute 6. Table 38 and fig. 48 contain an example of such results, obtained by exhausting the fog chamber at a stated time after the removal of radium. The drop in pressure is below the coronal fog limit of air and all precipitation takes place on ions. 60 20 nxl0~ 3 i ♦ N. FIG. 48 t *-5ec. *" + .6. 1 50 Fig. 48. — Decay of ionized nuclei (n per cubic centimeter) produced by radium in dust-free air, in the lapse of seconds. Table 38. Table 38. — Decay curve at dp =23. Radium on top. For the first five seconds 6 = 0.0019, for the first fifteen seconds 6 = 0.0022, etc., the values obtained ranging from 0.002 to 0.003. This is larger than the electrical datum 0.0014. Decay is more rapid than the equation warrants. Initial coronas are too large, final coronas too small, in spite of the presence of air nuclei, the number of which should be deducted at least in part. Other experiments show similar co- efficients. Thus the low curve of Miss Jos- lin would conform to 6=0.0023. Natu- rally the present method for 6 is much inferior to the electrical method, even if the two coefficients are identical ; but the 6 here is obtained under pos- sible complications with the larger gradations of the colloidal nuclei of dust-free air, though these are probably inefficient. If the values of i/n be inserted, the curves should be linear, since i/w = i/w + 6/, where t is the time dated since the occurrence of n . The line passing through the observations at 5, 30, 50 seconds is best adapted to represent the results, and compatibly therewith 6 -=0.0024 (« in thousands of nuclei per cubic centimeter) may be roughly assumed. These computed values of n are given in table 38 and shown in the chart (fig. 48). They are too low initially and too high finally, even if the air t. s. n x io- s . n x io- s . b =.0024 sec. 5-9 67 675 5 50 4i 37-3 10 4.6 32 25.8 15 4.0 21 19.7 20 3-5 15 159 30 3-3 11. 6 "•5 50 2.9 7-4 7-4 120 1-7 i-9 3-3 00 1.0 1. 1 96 VAPOR NUCLEI AND IONS. value is quite ignored; but the constant probably reproduces the true conditions better than the observation, remembering that the initial corona (t = o) is not quite invariable. A very important consequence may be deduced from these results. The equations specified may be written (?-)=»• Hence if the ratio of nucleations or of ions is known (for instance by the method of geometric sequences), n /n is given, and the absolute value of n may be computed if 6 is known. Now, if 6, for the case of ions, may be taken as identical with the value found in electrical experiments, where 6 = 0.0014, roughly and relative to ionization in thousands, bn = 0.0014 n\ where n\ is the true nucleation. Thus in table 39,6 = 0.0024, n = 67,500; therefore w' = (0.0024/0.0014) n Q or 115,000 nuclei per cubic centimeter. Quite generally, if n /n and b are determined from purely coronal measurements 6/0.0014 is the reduction factor for all the rela- tive nucleations to absolute values. Another very important consequence may be drawn. If the coeffi- cient is known from direct experiments, it will then be possible to standardize the residual curve (depressed asymptote) leading to the terminal corona. Thus if 6 = 0.0024 is roughly assumed, as an example derived from the data at dp = 24 cm. (table 37 and fig. 47), the value of the ordinates of the curve for dp = 31 would then be given by table 39 and fig. 49. Moreover, in any such curve, while the ordinates denote the computed number of ions, the abscissas denote the observed number of efficient colloidal nuclei and ions in the course of time, largely the former. Hence the curve gives an indication of the distribution of the precipitated water on the two groups of nuclei, different in size and present in different proportions, for the given supersaturation. Experiments of this kind are of the highest importance and the present cursory treatment is admitted provisionally, in view of a projected restandardization of the coronas of cloudy condensation, which the variety of results since obtained has made necessary. The curve for 6 = 0.0024 is shown in fig. 48; the two curves* i and n+i, i and w, in fig. 50. The initial descent of the graph for i and n is clearly steeper than would correspond to any exponential or hyperbola, and an equation of the form i (n+A) B = C is at least needed to express the data. The computation of the constants * Where i denotes the number of ions, n the number of nuclei per cubic centimeter. DECAY OF IONS. 97 would not be of any value. The table and chart show clearly enough how rapid a reduction of efficient nuclei is produced by the presence of but a few thousand ions. The results would have been much more striking if a more efficient form of apparatus had been used, since for dp = 31 cm. it is customary to obtain four or more times as many col- loidal nuclei in the absence of ions. Table 39. — Graduation of high pressure curve. i/rio =0.0133 = 0. 6=0.002; n =75X10-'; i/w = A =0.0133; b =0.0024. t. n Xio 3 atd/>« 3 i. a . hbt. n X io -3 . \ (n+i)~i } X io -3 = n X io _s . sec. 10 *29 0373 f26.8 2 20 32 0613 16.3 16 30 5i 0853 11. 7 39 40 62 1093 9.2 53 50 74 1333 7-5 67 60 85 1573 6.4 79 80 105 2053 4-9 100 100 120 2533 4.0 116 120 135 3013 33 132 140 140 3493 2.9 137 160 3973 2.5 180 5 145 4453 2.2 143 .0253s 39-5 15 .0493s 20.3 *(i+n)Xio-». t»Xio-». O 20 40 Fig. 49. — Decay of ions (f per cubic centimeter) in dust-free air, evidenced by increase of efficient colloidal nucleation (n per cubic centimeter). Table 39. 64. The same, continued. — In conclusion I may state that a number of experiments were made to test the rate at which the ions are gener- ated. If a is the number of ions produced by the rays per cubic centi- meter per second and b the coefficient of decay , dn/dt=a — bn 2 98 VAPOR NUCLEI AND IONS. This may be written 1 Cd (dn/dt) /fcj (dn/dtWa^ t = = I — : — : — - — -f C 2 VbJ (dn/dtWa- dn/dt from which, after integration, -2^/ab't+c N-y/a/b if n — n 1 . —2bn x t —Ac nXfit where A is a constant and w 1 = >/a/6, i+Ae- 2bn J x i-Az~ 2bn J n = n If t = 00 , n = n t , a = bn 2 , which is otherwise evident. The coefficient a is here taken as an absolute constant independent of n. In the pre- ceding paragraph it was roughly assumed that 6 = 0.002 if n is reckoned in thousands per cubic centimeter. Hence a = 0.000002 X n 2 . Thus if w = io 3 , a = 2 per cubic centimeter per second; if w = io 6 , a = 2 X io 6 , etc. Experiments were tried with the radium tube on a pendulum swinging above the fog chamber; also with the tube on an inclined plane moving rapidly across the chamber. But in both cases the results, which essentially require the opening of the stopcock of the fog chamber, are too involved to have any critical value, and they are therefore dis- carded here. 65. Condensation phenomena of the inclosed steam jet. Methods and results. — Some time ago (Bulletin No. 12, U. S. Weather Bureau, 1893), I obtained a series of results (shown for example in fig. 52) from observations of the behavior of the steam jet inclosed in a wide tube of thin sheet metal. The jet shown at / in fig. 50 plays into the tube AkA, about 2 inches wide and 2 or more feet long, the steam escaping at B. Sky-light L, from a mirror M, enters the tube axially, through a window a, and is observed through an opposite window, g. Room air enters at C, to cool the steam, and the temperature of the inflowing air is taken, as well as the pressure under which the steam escapes. Fig. 51 shows two such tubes arranged for differential work. These data are used in the construction of fig. 52, where air temperatures in degrees centigrade are horizontal and steam pressures in pounds vertical. STEAM JET. 99 It will be seen that for each temperature of the inflowing air there is a definite steam pressure at which the field of the tube just becomes opaque, and condensation within the jet, therefore, begins to be tumul- tuous. The edge of the opaque field is sharply marked, and above 15 it is possible to pass through it in a march of continually increasing steam pressures; for the opaque field vanishes with a kind of cusp, dependent for its position, naturally, on the apparatus used. What the figure imperfectly suggests, however, is the occurrence of similar loci of axial color, which run parallel to the edge of the opaque zone, on the clear side of it. The colors follow the spectrum series reversed, v, b, g, y, o, r, growing continually fainter and vanishing into daylight. The field becomes at once opaque if a strong nucleator like phosphorus is placed near C, fig. 50; or any color may be obtained in this way by carefully regulating the additions of nuclei, as shown elsewhere. Mere smokes, like salammoniac, are ineffective; in fact the field made opaque by phosphorus may be cleared by such smoke, added in reasonable quantity. Ordinary non-filtered air is practically ineffective. — This touches the first point to be made. I have shown elsewhere that the fog chamber and the steam jet mutually supplement each other; the former respond- ing measurably to nuclei reckoned in thousands per cubic centimeter, the latter to nuclei reckoned in millions per cubic centimeter, to speak roughly. In other words, the whole sequence of coronas, of which there are many periods visible in an apparatus of reasonable size, has been passed through when the occurrence of axial color begins, the latter end with particles so fine as to be optically ineffective. It is for this reason that ordinary non-filtered air, which produces such remarkable effects in the fog chamber, is almost without effect on the steam jet. A faint scarcely discernible pink tinge is all that is seen, and it is therefore possible to omit the filtration of air altogether. The use of dust-free air will not change the conditions, which lie wholly below the scope of the steam jet. This is even the case when the air is artificially dusted by less powerful nucleators like weak X-rays or weak radium, the additions being as a rule relatively insignificant. As the increase of steam pressure can only increase the supersatu- ration, it follows that the tumultuous precipitation of the steam jet characterizing the opaque zone must take place on the nuclei of dust-free air, for at this stage of the phenomenon the corona-producing dusts are ineffective. Much attention has been given by Professor Wood in this country, and by others abroad, to the occurrence of optical resonance in con- IOO VAPOR NUCLEI AND IONS. © © m m n [1 11 \\ M M' Fig. si- Fig. 50. — Section of inclosed steam jet and tube. Fig. 51. — Binocular inclosed steam jet. Fig. 52. — Chart showing margin of opaque zone for different steam pressures (pounds per square inch) and temperature of inflowing air. nection with similar color effects. But the case of the steam jet may be duplicated, in different ways, by the vapors of typical non-ionizing liquids, like gasoline, benzol, carbon bisulphide, etc., with even more saturated axial colors than are observable in steam. Electrical resonance can not, therefore, be effective here, where the fog globules are dielectric. The particles, moreover, are too large to fall within the lines of such an explanation. Returning to the diagram (fig. 52), it will be seen that at a temperature of about 40 there is no observable condensation; in other words, the steam passes through the tube like a gas, leaving the field quite clear. At 40 , therefore, the supersaturation at which condensation begins to take place on the nuclei of dust-free air is just reached; below it, at the given pressure, the supersaturation is in excess, and steam pressure STEAM JET. IOI must be relieved to reduce the excess before the field clears again, in greater measure as the temperature is lower, until at 9 steam issuing at ordinary pressure (without appreciable pressure above one atmos- phere) condenses spontaneously. The increments of supersaturation below 40 are in fact considerable. Thus, if the saturation at 40 be taken as the standard, these excesses would be roughly as shown in table 40. Table 40. — Estimated supersaturations. Temperature. . . Supersaturation Relation 40 35° 30° 25° 20° 15° IO° 12 21 28 34 3» 42 I 1.3 17 2.2 2.9 3-9 5-5 io -9 g/cm 3 . These values are to be added to whatever supersaturation preexists at 40 , seeing that the escaping steam is always suddenly cooled down from a temperature much above ioo°. In fact, the following data (table 41) may be adduced from the diagram, if t t be the temperature of the steam before, t 2 the temperature after, p t and p 2 the densities of the steam. Table 41. — Estimated supersaturations. h Po ft Pi 5 =Px/P 2 40° 35° 30° 25° 20° 15° IO° 50.9 39-3 30.1 22.8 17.2 12.8 9-3 112° 112° 112° 112° 111° 109 '05 894 894 894 894 865 811 7i5 17.6 22. 7 29.8 39-2 50.3 63-3 77 io- 8 g/cw 3 io~ e g/c*» 3 Nothing more than exhibition of the order of values is intended; but it will readily be seen that in comparison with Wilson's data (5 = 4.3 for rain-like, S = 7-6 for cloud-like condensation, and 5 = 9.9 for the sensitive tint), the supersaturations here are enormous, and that the condensations must take place on something approaching the molecular groups of the system, water — steam — air. It is therefore scarcely necessary to remark that the data are superior limits, particularly inasmuch as the influx temperature at C, fig. 50, and not the efflux B, were taken. But as the influx of air is swifter as the the steam pressure increases, the difference between these temperatures is not large. The escaping steam is always cool to the hand. It follows that the color data bear at once on the structure of dust-free air. The occurrence of definite loci for r, y, g, b, v, the fact that any color can be retained indefinitely, if (c&t. par.) the pressure-temperature 102 VAPOR NUCLEI AND IONS. conditions are fixed, shows that groups of nuclei counted by millions per cubic centimeter must be available for condensation long before the molecular sizes are approached. It is convenient to refer to the molec- ular aggregates in question by the term "colloidal nuclei." They are necessarily much smaller than ions. The number of such nuclei increases as the supersaturation is greater and the size needed therefore smaller until condensation is actually spontaneous on the molecules themselves. It is in this stage that one would naturally expect the fog particles of all sizes to produce a medium opaque to transmitted light, as is actually the case provided the fog particles are themselves large in comparison with the wave-length of light. It does not by any means follow, however, that the turbulent phenomenon is complete at the lower edge of the opaque zone. It is much more probable, seeing how gradually the violet drops into the opaque, as the supersaturations increase with increase of steam pressure at a given temperature of the inflowing air, that the number of nuclei continually increases. As this goes on (*. e., in a vertical march through the opaque zone) the time must arrive when the fog particles become small in comparison with the wave-length of light; for although the influx of steam has increased, the air influx increases in the same proportion. Under these circumstances, even though fog particles of widely different sizes are the rule, the Rayleigh effect of scattering is to be looked for. In fact, the yellows of the first order emerge from the opaque as mag- nificently saturated orange-browns and thereafter gradually become yellower. Deep crimson at the edge of the opaque does not appear. To prove that the passage from blue to yellow through opaque is due to an increase of nucleation, it is merely necessary to add phosphorus nuclei, when axial violet appears in the steam tube at a sufficiently high tem- perature. Without further change the bluish tone at once passes to yellow. Again, the intense oranges above the opaque may be obtained at once by introducing the fumes of the intensely nucleating sulphur flame into the steam tube. Finally, the probability of the Rayleigh effect is increased by the fact that colors of smaller wave-length than orange and yellow never occur above the opaque zone. 66. Summary. — The present chapter has made use of the method of testing the presence and number of relatively large nuclei by the depression produced in the terminal asymptote obtained when nucleation varies with the drop in pressure, or of the diminution in aperture of the terminal corona corresponding in a given apparatus to dust-free non- energized air. It was thus easily possible to trace the growth in number SUMMARY. I03 of the persistent nuclei produced by relatively intense X-radiation in the lapse of time, showing results similar to Chapter I, sections 14, 15. In the same way the occurrence of water nuclei in the presence of colloidal nuclei or of ions was clearly exhibited. These are due to the evaporation of small fog particles, until the increase of vapor pressure due to surface curvature is balanced by the decreased vapor pressure due to some independent phenomenon. Hence if the exhaustions proceed with a slightly opened filter cock, such nuclei must be present in greater number as the evaporation is faster, because increasingly more particles evaporate than subside. The result in the first place is a necessary alteration of aperture of successive coronas, all other conditions remaining the same, since the large ones evaporate more fog particles than those of lower order of size (periodicity of coronal diameter). In connection with this phenomenon it is particularly interesting to observe that not only the colloidal nuclei but the ions are available. Questions arise as to whether the fog par- ticles of positive and of negative ions evaporate to like water nuclei, what becomes of the charges, since it is not clear that they should be dissipated, what phenomena arise from the decreased mobility of the loaded ions, etc. The means developed in Chapter II were again applied to an effect produced (distance effect), when the radiating source is removed more and more from the fog chamber. The results are necessarily crude ; but they show that in case of pervious wood fog chambers the nucleations decrease more slowly than the first power of distance; they decrease faster for the glass fog chamber; faster still when the X-ray bulb is inclosed in a windowed lead case; fastest when the window is closed with a thin tin plate; but in no case do they reach the law of inverse squares. All this points clearly to the importance of secondary radiation in producing nuclei. The radiation arises both near the bulb and near the fog chamber, as well as in the region between. The number of nuclei produced is therefore dependent upon an integral extended over the whole interior surface of the room and throughout the intervening air. In case of exposure to weak radium (10 mg. 1 0,000 X) the asymptote of dust-free air is not reached until the intervening distance is above 150 cm.; and it is far from appearing after 1.5 cm. of lead have been pene- trated. The occurrence of minima of nucleation, when such weak radium is carried from a long distance quite up to the fog chamber, is again strik- ingly brought out, provided the drop of pressure is sufficiently above the fog limit of dust-free air. It is also observed when the ions are produced 104 VAPOR NUCLEI AND IONS. in the fog chamber by radium and decay in the lapse of time after the radium is suddenly removed. The decay curves all show minima for a drop of pressure above the fog limits of dust-free air; while the curve is normal for pressure differences below the fog limit. If the coefficient of decay for the ions within the fog chamber is the same as that found by purely electrical investigations, and if the relative numbers of nuclei corresponding in a given apparatus and given con- ditions of exhaustions are known, then a new method for the standard- ization of coronas in terms of the number of nuclei involved in their appearance, is suggested; for n = (n /n-i) bt if n Q and n are two successive nucleations observed under like conditions at t seconds apart and if b is the known coefficient of decay. Conversely, if n is given this method lends itself for the determination of nucleation ratios n /n, and for the experimental determination of the distribution of a given mass of precipitated water on groups of nuclei different in average size — one of the important of the outstanding problems. Finally, it was shown that data bearing on the graded character of the colloidal nuclei of dust-free air may be obtained most directly and throughout the widest range from the behavior of the inclosed steam jet. If (cat. par.) the steam pressure is intensified until the originally clear field (air dust being practically inactive), after passing a succession of colors, becomes clear again in consequence of absence of all con- densation, it follows that colloidal nuclei in vast numbers and of contin- ually decreasing size must successively become available for condensation until the fog particles are too small to be optically discernible. Each particular color corresponds to a group of colloidal nuclei characterized by their minimum size. It is still to be determined, however, whether these nuclei belong to the system of saturated vapor, or to the air, or to both. CHAPTER IV. DISTRIBUTION OF COLLOIDAL NUCLEI AND OF IONS IN MEDIA OTHER THAN AIR- WATER. COLLOIDAL NUCLEI AND IONS IN WET DUST-FREE CARBON DIOXIDE AND IN WET COAL GAS. 67. Apparatus. — The following experiments were made with the apparatus used in my last experiments* with dust-free air. The con- veyance tubes between the exhaustion chamber (5 feet long and 1 foot in diameter) and the condensation chamber (18 inches long and 5 inches in diameter) were about 18 inches long and 2 inches in diameter. The rapid exhaustion thus secured is effective, but has by no means reached a limit; there is still too much resistance in the connecting pipes. The data obtained with such an apparatus are comparable with each other, and nothing further than this is aimed at, since in view of the very high exhaustions needed, the constants for the computations of the absolute nucleations would in any case be lacking. It is a matter of convenience, however, to compute the data at the high exhaustion as if the conditions met with at the low exhaustion were indefinitely applicable, and this is the meaning to be given to n, the number of nuclei per cubic centimeter, in the present paper. Moreover, n refers to the nucleation left in the exhausted fog chamber, supposing that the nuclei are restored to the gas faster than they can be withdrawn by exhaustion or that the nucleation encountered is fixed for any definite environment. Otherwise n (to be multiplied by the volume expansion) would be very much larger. The carbon dioxide used was generated from calc spar and hydro- chloric acid. The gas was eventually passed through a solution of sodic hydrocarbonate, a long tube of dry bicarbonate of soda, a solution of silver nitrate, and a calcium chloride drying tube. Then it entered a filter (2 feet long), from which it was conveyed very slowly into the fog chamber. Coal gas taken from a gas pipe was treated in the same way. 68. Data for carbon dioxide. — In table 42, dp shows the drop of pres- sure on exhaustion, 5/30 (approximately) the angular diameters of the coronas when the eye is at 30 cm. in front and the source of light 250 cm. behind the fog chamber. The meaning of the other data is obvious, n being the nucleation. *Phys. Review, xxm, pp. 31-36, 1906. 105 io6 VAPOR NUCLEI AND IONS. Table 42. — Condensations in CO z generated from CaCo 3 with HC1. Gas washed as specified. dp. Corona. n xio- I. Washed in water, twice. 31.0 0) 31.0 10.7 31.0 7.8 31.0 6.4 31.0 5-3 31.0 2.7 31.0 2.4 31.0 2.0 19. 1 2. 1 16.9 2. 1 gyo wr g wyg wrg cor cor cor cor cor cor 410 260 140 100 60 8 5 2. 2. 2. II. Washed in solution of NaHCO- and AgN0 3 . 0) 10.7 7-9 3-3 2.4 2 8-4 'ii. 4 gyo w o g g|BP cor cor wrg wybg 410 305 150 14. 5 230 380 III. Same, with long tube of NaHCO,. 1.0 0) 1.0 II .2 1 .0 7-5 1 .0 4-3 1.0 4.4 gyo wog wB P cor cor 410 305 150 32 33 IV. Same, after 3 hours. 31.0 3 4 4 17.6 1 ? 17. 1 ? 17.7 1-2 19.4 1 ? 21.2 1-2 24.7 1-2 28.O 1 . 2 3I.O 2.0 3I.O 4 9.6 cor cor cor cor cor cor cor cor cor wrg 33 1 ?i 1 2 240 Corona. n x io -3 . V. AirbubblingthroughHCl,C0 2 ofT 31 12.0 12.0 wybg wybg 5 gyo 305 360 410 VI. Experiments with C0 2 repeated ( 6 ) 11 .0 6.4 4.2 ; 4-0 « 4 .2 2. 1 2.0 I 10. 3 4 1.8 4 3-2 5 6.7 7 9.8 gyo wo bg wrg cor cor cor wog green cor cor g'BP wr og 410 305 100 29 24 29 3 2 320 500 2, 12 140 270 VII. Same. 35o 36.8 38.0 44.0 yog g gBP BgBP 400 490 500 55o VIII. Gas ionized. X-ray bulb at D — 200 cm. 41.7 7-3 35-9 7-3 28.8 7-3 27.3 7-3 25.8 7-3 24.8 3-8 233 2.5 21 .7 1 . 1 24.2 4.2 26.0 6.8 25-5 5-6 wy'g wog wog wog wog cor cor Just seen cor wrg cor 155 146 130 125 120 20 5 1 25 105 62 1 * given for successive coronas beginning with air. 2 CO2 off; air inflowing. 3 Not yet air-free. * Exposure to X-rays. * No trace of HC1 gas escapes through the cleaning train. 6 8 given for successive coronas, begin- ning with air. 7 After 1 hour. 8 After 2 hours ; other data consecutive. NUCLEI IN WET C0 2 . 107 In parts I to III in the table the exhaustion (dp) is usually kept con- stant, with the object of observing the behavior of the mixture of gases, beginning with air and terminating with C0 2 . In part I about 8 exhaus- tions lead to the steady behavior of the latter gas. Inasmuch as the orig- inal air is still present to the extent of about i per cent, an admixture of this amount may be regarded as inappreciable. In the second part of the table, even four exhaustions, reducing the air content to about n per cent, nevertheless bring out the behavior of C0 2 . The same things happen less certainly in part III. In part II, however, on readmitting air, two exhaustions nearly suffice to restore the air corona. Parts III and IV show that coronas for a given mixture are reproduced after several hours. The relatively small coronas obtained with C0 2 as compared with air for the same drop in pressure and in the same apparatus gave rise to a suspicion that water nuclei associated with HC1 gas might be involved. In part V, therefore, atmospheric air was bubbled through HC1 and the gas then passed through the same drying train. After three exhaustions exceptionally high air coronas were obtained, showing that the method of producing the gas is of no consequence, thorough washing presupposed. Hence, in parts VI and VII of the table, the experiments with non- energized CO 2 were concluded. In part VIII of the table the gas is ionized by the X-rays with the bulb at a distance of 200 cm. The usual constancy of corona at high exhaus- tions is observed, while the fog limit is exceptionally high. 69. The behavior of carbon dioxide. — The graphs corresponding to table 42 are given in fig. 53, in connection with the corresponding data for air, both for the non-energized and energized states. It will be seen that in both cases the curves are essentially similar in their contours, but that the C0 2 curves require much higher exhaustion throughout than the air curves. In other words, the same coronas occur in C0 2 as in air, provided that in the former gas all the pressure differences are chosen to about 5 to 5J cm. higher than in the case of air. Similar relations hold for the fog limits, as was found directly by Wilson* in a different apparatus. The peculiar feature of these results is the degree of parallelism of the C0 2 and air lines, both for the non-energized and to a somewhat smaller extent for the energized state. Clearly the phenomena in both cases are alike in character, though lying far apart on the chart. 70. Cause of differences. — In view of the more coercible character of C0 2 one would naturally expect larger colloidal nuclei than in the case *C. T. R. Wilson: Phil. Trans. Royal Soc, vol. 189, p. 288, etc., 1897. io8 VAPOR NUCLEI AND IONS. 300 200 Fig. 53- — Efficient nucleation (n) observed in media water-air, water- C0 2 , water- coal gas, energized or non-energized at different exhaustions dp. of air, and it was with this anticipation that the data were investigated. Taken at their full value, however, they would point to a conclusion exactly the reverse of this. The colloidal nuclei in C0 2 are apparently smaller than in the air, and the same is true {cat. par.) for the ions. Unfortunately the precise meaning of these results is not clear, for in the first place the amount of adiabatic cooling may be written logr /r=(y-i)/y. log p /p and thus between two fixed pressures log — = const. The value of this fraction is for air, 0.29; for C0 2 , 0.22 ; for coal gas, 0.19. In other words, the amount of cooling is less in C0 2 than in air under like conditions, and hence the reduced efficiency of the fog chamber in the former case is qualitatively compatible with the thermal properties of CO 2 gas. Quantitatively, however, this compensation does not seem to be sufficient. For instance, the same corona is obtained in air and C0 2 when the pressure difference is 28 cm. and 33.5 cm., respectively. For NUCLEI IN WET C0 2 . 109 like nuclei this would imply the same degree of super saturation. Hence, if p x and p 2 be the vapor pressures of water before (20 C.) and after exhaustion, and p = j6 and p-dp be the corresponding pressures, the occurrence of like supersaturation implies that Pi / P-P2- *p y m Pi / p-p2-dp' \ Y P 2 ^ P-Pi ' P2^ P-Pi ' where p l and y refer to C0 2 and where y is the heat ratio for air. Hence y _ log (l ~ (dj/ + p 2 )/p) -log (l -Pilp) y log (1 - (dp +pj/p)-iog (1 -pjpy where the value for - = 1.10 as derived for direct experiment. y To compute the value of the same ratio from the coronal experiment it is necessary to know p 2 the vapor pressure on exhaustion and before condensation ensues. This datum is unavailable, but it must be greater than o and less than would correspond to the decidedly lower exhaustion dp = 17 (for instance) than the one applied (dp = 28). Hence limits of the value of *• may be computed by inserting p 2 = o and p 2 = o.2 respect- ively. The results are -, = 1.28, both for the superior and inferior limits, as would be otherwise evident. One may summarize these results as follows: Either the heat ratio of carbonic acid, y' t decreases in comparison with that of air, y, very rapidly y as temperature decreases, so that an average value of —=1.3 instead of - = 1.1 is to be used in the preceding experiments; or the colloidal Y nuclei in wet C0 2 , though distributed in a way closely recalling the case of air, are throughout smaller. 71. Nucleation increases subject to a uniform law of equilibrium. — The most interesting feature of the data for C0 2 is their repetition (under higher exhaustion) of the behavior of air. In other words, the two curves are closely parallel throughout their extent. This seems to imply that both are primarily dependent, not on supersaturation or cooling or on volume expansion, but on a common law of distribution of number with size of aggregates, such as is given by the theory of dissociation. In other words, a given drop of pressure of sufficient rapidity and from a specific initial value in each case generates the same number of colloidal nuclei, though it does not follow that the apparatus in all cases can entrap them. This is even true in different apparatus of different degrees of efficiency, as shown elsewhere. no VAPOR NUCLEI AND IONS. One might perhaps suppose that cohesion of molecules is more frequent when collisions are less frequent or that effectively more large particles reside in the exhausted gas. There are other cases given below which seem to suggest this peculiar inference. 72. Data for coal gas. — These are given in table 43 in the same way as in table 42, both for the non-energized and for the energized gas when the X-ray bulb is at a distance of 200 cm. from the fog chamber. The measurements were much less satisfactory than the above, the corona being thin and blurred. This may perhaps be due to the fact that mixed gases are under examination. Table 43. — Distributions of nuclei in coal gas, washed in water and AgN0 3 . dp. s. Corona. n Xio -3 . dp. s. Corona. nX io -3 . Coal gas 31.0 5-i cor 55 Coal gas 17.6 .0 .0 31.0 5.1 cor 55 19.9 i-7 cor 1.7 31.0 5-6 cor 70 31.0 4-3 cor 30 33-4 6-7 wyg(?) "5 X-rays. = 200 cm. 35-5 40.0 7.8 8.9 (Thin) wP(?) 160 250 44.1 9.2 w y' 45o Coal gas 40.1 6.6 w g 121 42.1 10.2 w y 45o 37-8 6.6 w g 120 Same. 39-2 yobg 340 35-7 6.6 w g 115 38.6 9-6 w b r 250 32.0 6.2 wrg 90 35-3 7-7 w P 200 30.2 5-3 cor 60 32.0 5-3 cor 60 28.6 4-7 cor 40 27-3 3-i cor 10 27.1 x 5-2 cor 50 25.8 3-o cor 9 25-7 4.6 cor 35 24-5 3-6 cor 16 24-3 3-4 cor 13 23.2 2.2 cor 3 22.9 2-3 cor 3-5 19.4 i-9 cor 2 21.7 i-5 cor 1-7 18.4 1.3 cor 1 20.6 l 2.3 cor 3-i 15.7 .0 19.4 .0 cor .0 1 Fluctuation of s not infrequent. 73. Character of the early results for coal gas. — As in case of C0 2 , the data for coal gas throughout lie in a region of relatively low pressure; i. e., large drops in pressure are needed to produce the coronas. Fog limits are correspondingly high. This will again be qualitatively in keep- ing with the low heat ratio 7. Apart from this, these first results with the hydrocarbon gas differ thoroughly from the character of the results for air and C0 2 . In the non- energized gas the nucleations rise to the pressure difference at a rapidly accelerated rate, and this continues to the highest value which dp applied NUCLEI IN WET COAL-GAS. Ill and long after the air and C0 2 nucleations have become stationary in the given apparatus. Similarly for the energized gas the increase of nucleation is very gradual and the asymptote is scarcely reached within the interval dp = 40 cm. of the experiment. All this is sharply in contrast with the rapidity with which air and C0 2 approach their respective asymptotes, as may be seen by inspection of the chart. Table 44. — Coal gas energized by X-rays at D = 10 cm. and in absence of rays. Two- inch pipes. Lead conveyance tubes. dp. s. Corona. n Xio~ 3 . dp. s. Corona. n Xio- 3 . 32.9 P) Slate bl. Bulb removed. 30.8 0) g'G 470 29.0 h Grayish 27-5 M Grayish 19.4 .0 0.0 26.0 11 w y ' 340 22.4 .0 21.7 5-7 w r ' 58 24.6 ? .0 20.6 3 5 cor 13 26.0 1 7 2.2 18.2 2 1 cor 2.3 27.9 3 8 21.5 16.9 .0 30.1 6 4 w y 97 24.0 10 5 wy 260 31.2 7 3 w be 137 22.7 7 8 g'B 125 32.8 9 5 wr 270 21.5 5 4 w r 50 34-8 10 4 w y 400 37-3 ?w y ' 410 Coronas too vague for measurement. 74. New data for coal gas. — In case of the above results the large vacuum chamber was not quite filled with coal gas and there may have been diffusion from one vessel to the other. Again, the rubber connecting- tubes of the first experiment (though in advance of the filter) were replaced by lead connecting-tubes, as rubber is pervious to coal gas. Whether from these causes or others, the results came out quite differ- ently after every part of the apparatus had been filled with coal gas. This is also shown in fig. 52 and the results themselves are inserted in table 44. The new results for coal gas now conform in character with the data for media of air-water and carbon dioxide-water, and for the non-ener- gized state, the coal-gas line lies between the curves for the other two media, as the figure shows. In like manner the curve for coal gas ener- gized by the X-rays now betrays nothing abnormal. In endeavoring to discover reasons for the differences between the data of table 43 and table 44, the presence of rubber tubing in the former case comes nearest to suggesting a solution. Carbon disulphide emits nuclei spontaneously; whether these may pass through the filter has not been tested ; but it is 112 VAPOR NUCLEI AND IONS. conceivable that the action of coal gas on vulcanized rubber may in a somewhat similar way to carbon disulphide be productive of nuclei. If these are larger than colloidal nuclei, though they may still be much smaller than ions, they would pass through the filter and give rise to the very phenomenon of depression of asymptote observed in table 43. 75. Conclusion. — Neither from the cases of C0 2 nor of coal gas does it follow that coercible gases have larger colloidal nuclei than non-coercible gases like air. The apparent results are the reverse of this. If the increased difficulties of condensation in the former are due to smaller heat ratios (y) as compared with air, there is no reason why C0 2 , where y is larger, should show greater condensation difficulties than coal gas where y is smaller. The distributions of nuclei suggest a common law of dissociation or chemical equilibrium. COLLOIDAL NUCLEI AND IONS IN DUST-FREE AIR SATURATED WITH ALCOHOL VAPOR. 76. Introductory. — In my report* on the solutional nucleus and else- wheref I came to the conclusion that the differences in promoting con- densation exhibited by positive and negative ions were more probably to be ascribed to the difference in chemical structure or composition involving a difference of size than to the electrical differences as such. Experiments made in Wilson 's apparatus by Dr. Donnan J with vapors of methyl and ethyl alcohol, carbon tetrachloride, carbon disulphide, benzol, and chlorobenzol show that the supersaturation needed to produce condensation was not necessarily greater in ionizing than in non-ionizing solvents. With similar apparatus Dr. K. Przibram§ recently examined a series of alcohols and other bodies ionized by the X-rays, obtaining among a variety of data a noteworthy result with a direct bearing on the question here at issue. It appears that whereas in the case of water vapor the negative ions are more efficient condensation nuclei than the positive ions, the reverse holds for the alcoholic vapors. In cases of methyl, ethyl, amyl, and heptyl alcohols (including some other bodies like chloroform) the positive ions invariably require less supersaturation to precipitate condensation than the negative ions of the same body. Interesting differences are therefore manifest in the behavior of vapors, and it seemed desirable to test the nucleation of a medium of * Structure of the Nucleus. Smithsonian Contrib., No. 1373, p. 161, 1903. t Ions and Nuclei, Nature, lxix, p. 103, 1903. X F. G. Donnan: Phil. Mag. (6), ill, pp. 305-310, 1902. 5 K. Przibram: Wien Sitzungsber., cxv, pt. ua, pp. 1-6, 1906. • ' NUCLEI IN ALCOHOL VAPOR. 113 ethyl alcohol and air in comparison with the media of water-air and water-carbon-dioxide hitherto examined. The former behaves in fact as if the nuclei were throughout larger than in the latter cases. 77. Apparatus and method. — The experiments were conducted with an apparatus in which the connecting pipes between the fog chamber (18 inches long, 5 inches in diameter) and the vacuum chamber (5 feet long, 1 foot in diameter) were 4 inches in diameter, containing a 4-inch counterpoised plug stopcock. The whole connecting system was about 22 inches long, one-half of it belonging to the fog chamber. Experiments made with water vapor, however, did not show any further marked advantage arising from the use of the large passage-way specified, over the former apparatus, in which the corresponding tube was 2 inches in diameter. It is therefore superfluous to adduce for comparison the new data for water vapor. The general method for work was that frequently described in connection with these investigations. 78. Properties of alcohol fog. — While the experiments of my preceding paper with the medium of water vapor and carbon dioxide gas showed unusually high values of the exhaustions needed to produce coronal condensation, the case of alcohol-air shows correspondingly low values of exhaustion as compared with those for water-air. The number of col- loidal nuclei entrapped by alcohol vapor are about 3.5 times larger than is the case for water vapor under like conditions. Hence the coronas for alcohol are exceedingly dense by contrast. They are also much less pure in color, and particularly at high exhaustions become fog-like. The phenomenon is coarsened and measurement less satisfactory. As the alcohol fog particles are larger in size, they subside more rapidly at the same exhaustion than water particles; but the occurrences are in the former case far from simple. While the corona (if not too large) remains nearly the same throughout the slow subsidence of water parti- cles, the corona for alcohol particles decreases one-half or more in size during this period. In other words, the alcohol particles experience very rapid growth during subsidence, from which it follows that many of them must evaporate to compensate in part for the eight-fold or more enlargement in bulk of the survivors. The same fact may account for the blurred coronas; for the true initial corona being very evanescent is probably not seen. Conformably with this view it is impossible to exceed large white-reddish forms in the present apparatus and to reach the high greens observed with water vapor. 79. Number of particles. — In order to determine the number of particles corresponding to a given corona, it is first necessary to compute the amount of alcohol precipitated per cubic centimeter of the exhausted ii4 VAPOR NUCLEI AND IONS. vessel by the sudden cooling incident upon exhaustion. This may be done by a straightforward approximation* with results shown in the following table, 45, where t t is the initial temperature of the saturated air within the fog chamber, t 2 the temperature after sudden exhaustion and before condensation, and t the temperature after the precipitation of the m grams of alcoholic fog per cubic centimeter. The drop in pressure is dp from p = 76 cm. at 20 C. The data of the last column will be presently explained. Tabl,k 45. — Precipitation of alcohol vapor at different exhaustions (dp), super- saturations (5), and radius (r), of nuclei. dp. 4- t 2 . t. m Xio 6 . S-fr/p+. r X io 7 . cm. 10 20 30 20 20 20 + 3-2 -15-3 -36.1 + 14.8 + 10. 2 + 3.2 Grams. 10. 18.3 22.8 2.42 6. 21 21.4 cm. 1.44 •75 .48 These may be compared with the case of water vapor. 8.5 17 22 30 20 20 20 20 + 5-8 - 9.6 -18.9 -36.1 + 14.7 + 8.8 + 4.6 - 3-5 2.6 4.6 5-5 6.4 2.17 5-68 69.7 165 •77 •58 •35 We may infer from the table that in a perfect apparatus and for true pressure differences! water fog particles would reach freezing (o° C.) at ^ = 24 cm. and alcohol fog particles at ^ = 34.5 cm. Moreover, for the same corona there must be on the average about 3.5 times more particles in the alcoholic fog than in the water fog, which accounts for the opaque- ness of the former. For the reasons adduced it is not worth while to express the results otherwise than in round numbers, for the data involved are inevitably crude. The assumption of the law of adiabatic cooling as far as - 3 6° C. is questionable, in view of the admixture of saturated vapor; but as the densities of vapor are for alcohol about 8 per cent that of air and for water vapor about 7 per cent, this approximation in a rarefied atmos- phere like the use of Boyle's law for a wet gas is probably admissible. It is different, however, with the latent heat of the vapor, which is required at the low temperatures, but is known (as a rule) only at temperatures near the boiling-point. From this and similar points of * C. T. R. Wilson: Phil. Trans., London, vol. 189, 1897, P- 300. f dp computed as p-p 2 in the way shown in Chapter II, and not the apparent value observed at the fog chamber. NUCLEI IN ALCOHOL VAPOR. 115 view, measurements of latent heat for the more common vapors at very low temperatures would be desirable. Finally, the point at which the drop in pressure ceases to be efficient, on account of the increasingly rapid inward radiation of heat from the vessel, is the most serious of the outstanding errors. I have endeavored to diminish it compatibly with the desideratum of a large and easily adjusted fog chamber by successively increasing the bore of the exhaust pipes and stopcocks; and this plan has been in the large measure suc- cessful. The extent to which the error is present, as the drop in pressure increases more and more, is nevertheless left unanswered. If the upper inflection of the distribution curve (fig. 56) is a criterion, i.e., the occur- rence of identical terminal coronas for successively increasing exhaus- tions, the fog chamber with water-air is efficient to about dp = 31 or 32 cm., with water and carbon dioxide to about dp = 37 cm., with alcohol and air to about dp = 20 cm. In the former case the vapor would be cooled from 20 to about — io° C. even after condensation; in the latter case to about +io°. On general principles and in view of the low temperature of the water, it would seem probable that the efficiency of the fog chamber must vanish gradually. But the appearance of the curves is such as if the action were unimpaired up to a given terminal drop in pressure. In every case the fog particles with the surrounding medium of vapor soon reach the temperature of the air again, so that additional moisture must arrive from somewhere. It is remarkable that the marked con- stancy of the water coronas in perfectly tight apparatus during this period gives no evidence of the evaporation; while the alcohol coronas decrease one-half in aperture or the preponderating fog particles actually grow. Even if this is compatible with the evaporation of the smaller particles, there is again no evidence for it. Much of the moisture must therefore come from the wet cloth and the water within the vessel, which are not cooled by the expansion. 80. Size of the nuclei. — Here it may be worth while to inquire into the reason why the precipitation in alcohol is apparently so much easier; or, what is the same thing, into the estimated size of the nucleus on which precipitation takes place in these several cases. The Kelvin equation as modified by Helmholtz* may be used for this purpose, as was done by the latter and by Wilson t in the form p T /p M =e 2 T/Rsdr where p r and p M are the vapor pressures at the convex areas of radius r and radius infinity, respectively, T the surface tension of the liquid of density (s), R the gas constant of its vapor at the absolute temperature (#) . * Helmholtz: Weid. Ann., xxvn, p. 524, 1886. f Wilson: Phil. Trans., vol. 189, p. 305, 1897. n6 VAPOR NUCLEI AND IONS. Fig. 54. — Supersaturations (S) of media air-water and alcohol -water, and estimated radius (r cm.) of smallest efficient nuclei, at different exhaus- tions (dp), m, molecular radius. Table 45. Since p T is the adiabatically reduced vapor pressure (with- out condensation) in the vol- ume expansion due to the drop of pressure (dp) and p M the normal vapor pressure at the same temperature (?9 = 2 7 3° + *2 m table 1), r follows from the equation. The values of p r lp w and r so found are both given in table 45, and have been constructed in the chart (fig. 54), where their relation to the usual order of molecular size is also indi- cated. Clearly these values of r, the radius of the nuclei differing so little from molec- ular radii (say io~ 8 ), can only indicate an order of values; for apart from the difficulties above enumerated in com- puting #, r depends on sur- face tension (T), which has no meaning for molecular dimen- sions. Granting this, it is none the less remarkable that the values of r obtained should be so nearly alike for water and alcohol where different constants (T, R, s, etc.) occur throughout; in other words, that at a given temperature a given drop of pressure will condense both vapors on nuclei of about the same size. In so far as these estimates are admissible it follows that the alcohol- air nucleus is larger than the water-air nucleus, since in the former case coronal condensation begins at about dp = i$ cm. where r = io~ 7 cm. and in water vapor it begins at dp = 26 cm. where r= 4 X io~ 8 cm. about, less than half as large. These relations once established are retained through all successions of nuclei, as the following data for alcohol vapor in comparison with water vapor show. It is a little difficult to under- stand why the ionized nuclei in alcohol vapor should, like the colloidal nuclei, be larger than the corresponding cases for water vapor, unless the ions are aggregated vapor nuclei, a point of view tentatively advanced elsewhere; but this larger alcohol nucleus suggests that it is primarily NUCLEI IN ALCOHOL VAPOR. II 7 Table 46. — Colloidal nuclei and ions in alcohol vapor, energized or not; 4-inch cock. dp. s. Corona. n X10-*. I. April 23. Not energized 30.8 28.4 10 i wr > 1,060 10 i wr ' 1,005 26.0 10 i w r ' 950 23.0 11 1 W y' 1,100 21. 1 10 'w c 615 19. 1 7-7 gBP 385 17-3 4.8 cor 108 16.0 2.4 cor 11 4-7 1-5 cor 4 13-4 1.1 cor 3 12.3 (?) ( 2 ) 10.2 (?) ( 3 ) 19.9 (?) (') 12.7 ( 2 ) 16.7 5-o cor "5 18.5 7-7 gBP 380 20.0 10.5 w r' 680 I 1 , April 24 34-0 42.0 10 x wr' 940 1,040 10 *wr' 38.8 11 2 w 0' 1,170 34- 2 11 *w 0' 1,110 32.2 11 'wo' 1,080 29-3 12 'w y' 1,270 26.8 'wy' 1,210 24.0 l w y' 1,120 21.9 11 'wr' 720 20.0 10 l w c 600 18.3 8.2 wy 340 16.7 3-5 cor 38 15.0 1.0 cor 3 Later 14.8 1 cor 3 16.3 5-o 3 wr' "5 18.3 7-6 w y b 320 19-5 9.0 w r' 585 21.7 10.5 w r' 720 22.0 10.5 w r' 740 April 25 26.0 12 w y' wy' w y' 1,190 1,160 I XI. X-rays on. D •=• 15 cm 25.2 17.7 4 I2 12 920 151 II w r' 650 14.0 8.4 w y B 275 12.9 2.6 cor 13 11. 5 •5 (?) 11. 5 .O (?) X-rays off 11 .0 5 .o IV. X-rays on. D = 15 cm 12.4 13.2 5 .o 5.1 cor IOI 14.2 8.1 w B P 275 151 10. w r' 490 16. 1 10.5 w r' 580 April 26 36.6 11 w y' 1,420 1 Accompanied by dense fogs, rapidly sub- siding, with coronas dwindling. 1 Small coronas probably due to shaking. 8 All coronas accompanied by dense fogs. * Coronas decreased in size during sub- sidence, more than one-half. •Coronas often due to solutional nuclei; produced by shaking. n8 VAPOR NUCLEI AND IONS. 30 16 18 20 22 24 26 28 Fig. 55. — Aperture of coronas (s), observed at different exhaustions (dp), in media of alcohol-air, water-air, energized or non-energized. Table 46. .y FIG. 56 ggarr" t\ *** Jfj, / / ALCOHi ■>L -AIR /i 1 H if 1 oA WATER 3 TlYlO' 3 A ,4 \ 1 / /o " £- — */ #/ -t& A / +*p 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 Fig. 56. — Efficient nucleation n in energized or non -energized media of alcohol-air and water-air, at different exhaustions {dp). Table 46. the saturated vapor which furnishes the colloidal nuclei and that the gas is only secondarily involved. 81. Data for alcohol vapor. — The behavior of alcohol vapor is shown in the usual way in table 46, where dp is the sudden fall in pressure from atmospheric pressure producing the corona of the angular diameter 5/30, when the eye and the source of light are at distances 40 and 250 cm. on the opposed sides of the fog chamber. The nucleation n is com- puted in the way given above, section 79, and indicates the number of NUCLEI IN ALCOHOL VAPOR. 119 nuclei in the exhausted fog chamber. It is assumed, therefore, that the nuclei are reproduced more quickly than they can be withdrawn by the exhaustion. Measurement of j is not very satisfactory, as the coronas are blurred and accompanied by dense fogs and changes rapidly on subsidence. The effect of X-radiation is shown in parts III and IV of the table leading to the same terminal corona as for the non-energized vapor. The results have been given graphically in figs. 55 and 56, the former referring to coronas (s), the latter to nuclei (w), in connection with earlier results for media of water-air and water-carbon dioxide, the same con- densation apparatus and method underlying all experiments. One may notice at once that in the cases C0 2 -water and air-water, both for the non-energized and for the energized state, the observed data would be obtained by shifting the air-water diagram as a whole to the right, as if the cooling in case of C0 2 -water vapor were less efficient. The graphs are of the same kind, nearly parallel, and all of them (energized or not) terminate in the same asymptote or large green-blue-purple corona. The alcohol curves are different from these chiefly in three respects: (1) Though the graphs both for the non-energized and energized states again terminate in the same asymptote, this is not the green-blue-purple corona, but the white-yellow corona which lies slightly below it; (2) the curves as a whole lie with a somewhat larger slope in a region of much lower exhaustions (dp); (3) the number of nuclei caught in alcohol vapor is relatively very large. The second and third observations have already been discussed. The first deserves especial consideration. The question occurs at once why both the energized and the non-energized curves should terminate in the same final corona, irrespective of the size of the nuclei, and why this should be lower for alcohol than for water. For the ionized state one might infer that the total number of ions has been precipitated, as is actually the case for low ionization; but if for strong ionization this were true for alcohol vapor it could not be true for water vapor where the number of ions caught is less than one-half the number in alcohol. In general, it is improbable that the terminal corona for ions should in such a case be the same as the terminal corona for colloidal nuclei. The explanation which seems plausible to me is this: Each nucleus must drain the air of its supersaturated moisture within a certain radius, large as compared with the size of the nucleus and increasing in the lapse of time. A limit of the phenomenon will be reached when for an indefinite number of graded nuclei the enveloping spheres free from supersatu- 120 VAPOR NUCLEI AND IONS. ration form a system in contact. In case of water vapor the distance between centers would be 0.014 cm.; in case of alcohol 0.010 cm., dis- tances which are both enormous as compared with the estimated size of nuclei (r, table 45, fig. 54). The greater distance which belongs to water vapor would be in keeping with its greater diffusivity, though this surmise does not work out for water-carbon dioxide as compared with water- air. At all events, when the limiting number of nuclei has been captured, the apparatus is powerless to produce condensation on a greater number of nuclei, be they relatively large as the ions or small as the colloidal nuclei, however many other inefficient nuclei may be present. ABSENCE OF COLLOIDAL NUCLEI IN STRONG ODORS. 82. Introductory. — Throughout the course of my work I have been endeavoring to find whether bodies with strong odors and presumably large molecules could be regarded as a source of colloidal nuclei; or whether there is any relation between the colloidal nucleus and the odors. In case of carbon disulphide, from which nuclei apparently escape spontaneously, this would seem to be the case; but it is yet to be proved that vapor of carbon disulphide, if carefully filtered, neverthe- less still produces colloidal nuclei in the fog chamber. In the present instance it is not improbable, if relatively few CS 2 molecules are oxidized SO 3 , that sulphuric acid nuclei are the result. The experiment which has already been discussed is well worth while ; but I did not make it, in the fear of contaminating the vacuum chamber. The following experi- ments show, however, that even in the extreme cases odors are due to molecules, and that in relation to the fog chamber, apparently large mole- cules are still quite negligibly small in comparison with the colloidal nuclei. 83. Data for camphor, turpentine, naphthalene. — The results obtained are given in table 47 in the usual way. The glass fog chamber was provided with a hole in the (thick) bottom, about 2 cm. in diameter and closed with a rubber cork. It was free from leakage from without. The bodies to be examined were introduced through the hole in question, and suspended in the middle of the fog chamber by aid of a wire-gauze tube 20 cm. long and 2 cm. in diameter. After putting this tube in place the air was carefully cleansed by precipitation of dust. On May 5, 6, and 7, after putting the apparatus together, the medium within was first examined without introducing the odoriferous body. The internal sources of spurious nucleation gradually vanish. The NUCLEI IN STRONG ODORS. 121 Table 47. — Miscellaneous experiments. Water nuclei effect of metal surfaces, of odors, etc. dp =/> - p 3 . Four-inch pipes and plug cock. Date, etc. dp. s. Corona. n X 10- 8 . 30.8 9.4 wrg' 240 30.6 '6.5 cor 105 30.8 IO.O wrg' 260 30.8 12.0 wy bg 380 30.8 *ll.O wrg 275 30.8 II .0 wrg 275 30.5 10.0 wpg 230 30.6 II .0 wy'g 380 30.7 II wrg 305 30.6 II wrg 275 30.8 1 1 wog 305 30.8 II wyg 380 30.8 13 g' to gy 410 30.8 13 g' 455 30.8 13 g' to gy 425 30.6 13 g' to gy 425 30.6 12 yo 410 30.7 ( 3 ) gyo 410 30.6 13 gtogy 425 30.6 13 gBP 455 30.8 13 gBP 455 May 5. Wet cloth lining in exhaust tube May 6 May 6 May 7 Examination of odors: 3 May 8. Air only May 9. Air, camphor in May 10. Camphor May 11. Camphor Air without camphor May 12. Air May 13. Turpentine May 15. Naphthalene 4 , 1 Foreign nuclei not absent in second exhaustion. 2 Hole drilled in bottom of glass fog chamber, closed by rubber cork. 3 Apparatus blurred. 4 Enlargement of coronas due to second (smaller) exhaustions. exhaust pipes were cloth-lined and the cloth kept wet to guard against minor saturation error. On May 8, with the introduction of camphor, the nucleation apparently rises to a maximum (g-b-p. corona) ; but the same result is maintained after the removal of the camphor and the gauze (tested in many observations not recorded in the tables). Hence the camphor nuclei can not be larger than the colloidal nuclei of air- water; for the presence or absence of camphor within a fog chamber is inappreciable. The same conclusion follows from the experiments with turpentine (where the walls of the fog chamber were speedily blurred from distillation of small quantities out of the lamp wick holding the liquid) and from naphthalene. It is fair to conclude that presumably large molecules are nevertheless bodies of an inferior order of size rela- tive to the colloidal nuclei of dust-free air. 84. Summary. — Media of coal gas and water require higher exhaustion than media of air and water, media of carbon dioxide and water higher than coal gas and water, to precipitate condensation in like degree, 122 VAPOR NUCLEI AND IONS. other things being equal. On the other hand, media of alcohol and water require smaller exhaustion to precipitate an alcohol fog in corresponding degree. The general character of the phenomena in all these several cases are, however, the same. Their relation to each other when the media are energized in like manner by the X-rays is also the same. (Pf. fig. 56.) It does not follow, therefore, that wet coercible gases like carbon dioxide and coal gas contain larger colloidal nuclei than wet air. The apparent result would be quite the reverse of this. Neither can the difference be explained in terms of the respective value of the ratio of specific heats (y); for in relation to condensation, the order is air, coal gas, carbon dioxide; in relation to y, air, carbon dioxide, coal gas. The probable occurrence of large colloidal nuclei in media of alcohol and air and relatively small colloidal nuclei in water and air seems to show that the colloidal nuclei are primarily to be associated with the saturated vapor, and that the gas involved is of secondary importance. It can not be a question of mere solubility of gas in the vapor, for this is strongest in water and carbon dioxide, in which there is no evidence of large nuclei. The curves of distribution of number with size are again such as to suggest a common law of equilibrium. The presence of strong odors (camphor, naphthalene, turpentine, etc.) in the fog chamber is without an appreciable effect. Hence the col- loidal nuclei can not be large molecules merely, and colloidal nuclei can not be ascribed to chemical impurities in the gas. Continued slow exhaustion seems to be favorable to the growth of fog particles and the formation of rain. Both for the energized and for the non-energized state, the graphs for the same media, though referring to nuclei of widely differing sizes (ions, colloidal nuclei), again terminate in the same asymptote, or end in the same terminal corona. This is not identical for different media, however, the corona being of lower order though denser for alcohol and air, of higher order but thinner for water and air, for instance. The number of nuclei caught {cat. par) in alcohol and air is much larger than in water and air. When the limiting number of nuclei has been reached by condensation, the system is powerless (in a given apparatus) to con- dense on a greater number of nuclei, be they relatively large like the ions or small like the colloidal nuclei. CHAPTER V. THE COTEMPORANEOUS VARIATIONS OF THE NUCLEATION AND THE IONIZATION OF THE ATMOSPHERE OF PROVIDENCE. By Lui 60 v n s ' 1 f "^^ ^ ^^^i r""^ - H " : • <0 20' xW 3 _ ; ) f» Positive ions Negative ions Nuclei Aug. Sept. Oct. No, Dec. Feb. Mar. Fig. 62. — Average monthly ionization (thousands per cubic centimeter), positive and negative as stated, and nucleations (thousands per cubic centimeter), observed between August, 1905, and March, 1906. the new curve, the absence of maxima in December, are striking. One may note the upward march of the successive curves 1902-03, 1903-04, 1904-05. 9 1 . Measurement of ionization. — To determine the number of ions in the atmosphere, Ebert's* well-known apparatus was used. This consists of a tubular condenser, the inner coat of which is charged and in con- tact with a graduated electroscope. The air to be examined is passed through the condenser by an aspirator-fan propelled by clockwork. The air delivery of the machine is also carefully standardized. In order to test the same air which yielded the nuclei for the preceding measurement, the electrical apparatus (E, fig. 576) was swung from the outside of a window on a long swivel bracket (B) . In this way it could be drawn near the window for charging and examination with appropriate * H. Ebert: Illus. Aeronaut. Mittheilungen, October, 1902, pp. 1-10. 142 VAPOR NUCLEI AND IONS. lenses or moved to a reasonable distance away from the window during the passage of the air to be tested. In winter all measurements must be made with a galvano- scope on the outside of the house. Thermome- ters are shown at T. The difficulties encoun- tered in using this ap- paratus in cold weather will be investigated later (section 94). Here some reference to its constants is in place. The quantity of air passed through the condenser in the fiducial time (about 10 minutes) was 1.0357 X io 8 cubic centimeters; the capacity of the condenser 17.74 Oct. Nov. Dec. Jan. Feb. Mar. Apr. May June Ju/y Aug. Sept. cm. Hence if V is the Fig. 63. -Atmospheric nucleation (thousands per cubic dro P of potential in volts centimeter) from October, 1902, to March, 1906. during the fiducial time specified, 0-7.7 v/3 °°. ~ v - . io 357Xio 17.52X10 denotes the charge in 1 cubic centimeter of air. As 3.4 X io~ 10 is the electrostatic charge per electron, 70 60 FIG. 63 t^iti' ATM, "Di NUCL /St" El 50 06 i9° s 30 20 10 / 1 \ P<. 1 $| S> t r 3 % ^ L -»-r nXio _3 = Q 3-4XIO-' — , nearly, shows the number of ions per cubic centimeter. Measurements to find their velocity were not made, as this would have carried me too far from the purposes of this paper. I may add that a similar apparatus was installed, in which the winter indraft of cold air (through a condenser), measured by an anemometer, was utilized. The object here was to determine the hourly variations of ionization. Though many observations were taken, their meaning is vitiated by the temperature discrepancy mentioned in section 94. For this reason, perhaps, a periodicity similar to the one discovered by Wood and Campbell (Nature, April, 1906, p. 583) with stagnant air was not detected. ATMOSPHERIC IONS. 143 92. Data for ionization. — In table 51 the date and hour are given in the first and second columns. The fall of potential, V-v (v being the fall in the absence of the aspirator air current), in volts, during the ten minutes of observation, is shown in the fourth column, marked "volts." From this the charge (Q) in electrostatic units per cubic meter, and the number of ions (w) in thousands per cubic centimeter, are computed. The sign of the charges is indicated in each case, and q denotes their ratio (Q+/Q-). The correction (v) is regarded as negligible. Table 5 Date. Time. Volts. Q. nXio- 3 . Date. Time. Volts. Q. nxio~ s . 1905. 1905. July 26 4.0 3-7 +0.21 0.62 Aug. 9 9.2 10.9 + 0.62 1.82 4.6 + .26 .76 1.7 - . 10 .29 5-1 + .29 .85 8.2 + -47 1.38 7-8 + • 45 1.32 5-2 - 30 .88 6.1 + • 35 1.03 6-3 + .36 1.05 5 5-9 + • 34 1 .00 4.2 - .24 .70 29 9.2 50 + .29 .85 6-3 + .36 1.05 5-9 - • 34 1. 00 4.2 - .24 .70 4.0 6-3 + • 36 1 .02 12.3 5-2 + .30 .88 4.2 - • 24 .70 6.3 - .36 1.05 3i 10. 7-7 + .44 1.29 10 9-5 8.2 + .47 1.38 3-o - .18 •52 5-2 - 30 .88 3-o 7.8 + • 45 1.32 12 7.2 + .41 1.20 59 - • 34 1 .00 7.2 - .41 1 .20 Aug. 1 10. 2.6 + • 15 •44 11 10 6.3 + .36 105 2.6 - 15 •44 5-2 - 30 .88 3-o 6-3 + .36 1 .02 12.3 8.2 + .47 1.38 8.1 - .46 i-35 5-2 - .30 .88 2 9.2 6.9 + • 39 1. 14 12 9.0 6.3 + .36 105 56 - • 32 •94 6.3 - 36 1.05 4.0 5-9 + • 34 1 .00 12.0 5-2 + .30 .88 3-i - .18 •52 6.3 - .36 105 3 10. 8.2 + .47 1.38 14 9-4 11. 9 + .68 2.00 5-5 - .31 • 91 II. - 63 i-85 4-3 6.7 + .38 1. 11 II.3 13.8 + .79 2.32 4-4 - .25 •73 II. - -63 1.85 4 10. 5-3 + • 30 .88 12.3 II. + .63 1.85 4-4 - • 25 •73 II. - 63 1.85 43 2.6 + 15 •44 15 10 10 + -57 1.67 4.8 - • 27 •79 8.2 - -47 1.38 5 10. 4.2 + • 24 .70 10. + -57 1.67 5-3 - • 30 .88 8.2 - -47 1.38 4.0 4.6 6-5 + .26 • 37 .76 1.08 9.1 27.8 + -52 1 -52 7 9-4 10. + • 57 1.67 12.3 91 + -52 152 6-7 - .38 1 . 11 91 - .52 1.52 12.3 9.0 + .51 1.50 16 10 8.2 + -47 1.38 9-5 - • 54 1.58 5-2 - 30 .88 8 10. 5-5 + • 31 •9i 1 .0 10. + -57 1.67 3-2 - .18 •52 7.2 - -4i 1.20 9 9.2 6-3 + • 36 i. 05 17 10. 8.2 + -47 1.38 5-5 • 31 .91 5-2 - 30 .88 144 VAPOR NUCLEI AND IONS. Tabus 51.— Continued. Date. Time. Volts. Q. n x io -3 . Date. Time. Volts. 2- n X 10- 3 . 1905. 1905. Aug. 17 12.3 10. + c •57 1.67 Sept. 2 9-7 3-7 - 0. 21 0.61 10. - •57 1.67 i-4 6.0 + •34 1. 00 18 10. 8.2 + •47 1.38 2. 2 - . 12 •37 9.1 - •52 152 4 10.5 5-i + .29 .85 12.0 9-i + •52 1 52 2.8 - .16 •47 9-i - •52 152 i-7 5-4 + •31 .91 19 9.0 10. + •57 1.67 10.4 - • 59 1.74 9.1 - •52 152 7 10.3 4.8 + .27 .80 12.0 3-3 5-3 5-4 + •30 •3i .88 •9i 21 10. '8.2 + •47 1^38 4.2 - .24 .70 9.1 - •52 152 8 10. 2 7-i + .40 1. 18 1.0 7.2 + .41 1 .20 6-5 - •37 1.09 2. I - . 12 •35 2-5 8.6 + •49 1.44 22 10 6-3 + .36 1.05 9-5 - • 54 i-59 6-3 - .36 1.05 9 9-7 50 + .28 .84 1 .0 3- 1 + .18 •52 6.4 - .36 1 .06 4.2 - .24 .70 2.2 8.9 + • 5i 1.50 23 9-3 10. + •57 1.67 4.2 - .24 .70 8.2 - •47 1.38 11 9-9 3-7 + . 21 .61 12.0 7.2 + .41 1 .20 4-i - .23 .69 10. - • 57 1.67 2.4 2-3 + .13 .38 24 10. 6-3 + .36 105 2.8 - .16 •47 6-3 - •36 1.05 12 10. 1 10.8 + .61 1.79 12.0 4.2 + •24 .70 1.9 - . 11 •32 5-2 - •30 .88 2.7 2.2 + . 12 .36 25 10. 5-2 + .30 .88 3-3 - •19 .56 31 - .18 •52 13 10.3 4-8 + •27 •79 1 .0 8.2 + •47 1.38 4.0 - .23 .68 6.3 - .36 1.05 2-3 4.2 + •24 .70 26 9.0 8.2 + •47 1.38 30 - • 17 •50 6.3 - .36 1.05 14 10.4 9.2 + •52 1-53 12.0 9.1 + • 52 152 6.7 - .38 1 . 12 8.2 - •47 1.38 2-5 8.2 + •47 i-38 28 10.4 6.2 - •35 1.05 7-6 - •43 125 3-4 + .19 •56 15 10.4 6.0 + •34 1. 00 2.7 2.4 - H .41 7-7 — .44 1.29 3-2 + .18 •54 2-5 10. 1 + • 57 1.68 29 9-9 3-7 + . 21 .63 3-5 - .20 •59 3-7 - . 21 .63 16 10.5 4.2 + •24 .70 2.6 5-o + .29 .86 2.4 - •H .41 4-5 - • 25 .76 2.0 8-3 + •47 1 .40 30 10.5 4.8 + .27 •79 24 - •14 •41 30 - •17 .48 18 10. 2.8 + .16 •47 2.6 4-5 + .25 .76 5-3 - •30 .88 5-i - .29 .85 3-5 8-3 + •47 1.40 3i 10. 3-o + • 17 • 50 2.8 - .16 •47 5-6 - 32 •94 19 9.8 7-7 + •44 1.29 2.7 5-9 + •34 1. 00 1.8 - . 10 •3i 4.2 - •24 •7i 3-3 3-3 + • 19 .56 Sept. 1 10.5 5-4 + •3i •9i 2-3 - .13 .38 4-7 - .27 • 79 20 10.2 3-5 + .20 •59 3 6.8 + •39 1. 14 1 . 2 - .07 .20 •4 - •23 .67 3-7 4-7 + .27 •79 2 9-7 5-i + .29 .85 4.2 •24 .70 ATMOSPHERIC IONS. M5 Table 51.- -Continued. Date. Time. Volts. Q. n X 10- 8 . Date. Time. Volts. Q. n X io- s . 1905- 1905. Sept. 21 8.7 8.3 + •47 1.40 Oct. 7 2.6 3-5 -0 .20 o.59 4.2 - .24 .70 9 12 1 8.7 + 50 1 .46 3-5 4-5 + .26 .76 7.2 - • 41 1.20 4.6 - .26 •76 4-5 16.6 + • 95 2-79 22 9.8 1.9 + . 11 •32 2-4 - • 14 •4i 2.6 - • 15 •44 10 9-9 2.8 + .16 •47 50 2.8 + .16 •47 4.8 - .27 •79 2.4 - •14 .41 2-5 4.2 + 24 .70 23 9.0 7-8 + 44 1.30 8.1 - .46 1-35 7-7 - •44 1.29 1 1 9-8 6.0 + .34 1 .00 25 11. 8-3 + 47 1.40 8.9 - 51 150 95 - •54 i-59 2. 2 2-3 + 13 •38 2.7 5-3 + 30 .88 4.2 - • 24 .70 10. 1 - .58 1. 71 12 9.0 3-4 + .19 •57 26 9.0 4-5 + .26 .76 8.9 - .51 150 4.1 - 23 .69 2.7 6-3 + • 36 1 .02 3-7 6.8 + •39 1. 14 7-7 - • 44 1.29 9-3 - •53 1.56 13 11. 3-7 + .21 .61 27 9.0 4-5 + .26 .76 8-9 - • 51 1.50 2.7 - • 15 .46 2.6 1.4 + .08 •23 5-5 1 . 1 + .06 •19 3-o - • 17 •48 3-5 - .20 • 59 14 10. 7.2 + .41 1 .20 28 9-5 3-7 4- .21 .61 7-i - .40 1. 18 4.8 - .27 • 79 2.6 4-9 + .28 .82 3-4 8.3 + •47 1.40 4.2 — • 24 •70 5-4 - •3i •9i 16 2-5 5-9 + • 34 1. 00 29 9.2 7.0 + .40 1.17 3-7 - .21 .61 6.8 - •39 1. 14 4-7 M + .08 •23 2-5 2.4 + •14 .41 2.4 - .14 .41 1.8 - . 10 •31 17 10.3 6.1 + 35 1.03 30 9.0 1.9 + . 11 •32 7-4 - • 42 1.23 30 - .17 .48 2.7 7-5 + • 43 1.26 3-2 3-7 + .21 .61 3-5 - .20 •59 1.4 - .08 .23 18 9.2 30 + • 17 .48 Oct. 2 9-i 2-3 + •13 .38 8.0 - .46 1-35 24 - .14 .41 2.7 2.4 + • 14 •4i 2.7 1.8 + .10 •31 2-4 l - .14 .41 2.4 - .14 •4i 19 12 4.2 + • 24 .70 3 10. 3-6 + .20 .60 7-4 - 42 1.23 30 - 17 •50 3-5 6.7 + • 38 1 . 12 2.8 7-9 + •45 131 4.8 - 27 •79 4-8 - .27 • 79 20 11. 6 6.4 + • 36 1.07 4 9.8 36 + .20 .60 2.8 - .16 •47 4.2 - .24 .70 2.6 6-5 + .37 1.09 2-9 3-o + •17 •48 1.9 - . 11 •32 3-6 - .20 .60 23 12.2 4-2 + • 24 .70 5 10.5 24 + .14 •41 6.9 - • 39 1. 14 6.0 - •34 1. 00 4-3 6.4 + .36 1.07 6 9.2 6.6 + •38 1. 11 7-7 - • 44 1 .29 9-5 - •54 i-59 24 10.2 i-7 + .10 29 31 7.2 + .41 1 .20 3-3 - .19 56 10.4 - •59 i-74 2.7 3-1 + .18 •52 7 9.2 6.5 + •37 1 .09 1.8 - . 10 •3i 6.6 - •38 1. 11 25 12. 1 6.2 + • 35 1.05 2.6 3-6 + . 20 .60 8.2 * • 47 1-38 146 VAPOR NUCLEI AND IONS. Table 5 1 . —Continued. Date. Time. Volts. Q- WXIO- 3 . Date. Time. Volts. Q. n X io~ 3 . 1905. 1905- Oct. 25 4.0 i-4 + 0.08 0.23 Nov. 14 10.7 5-4 -c ►•31 •9i 3-6 - . 20 .60 16 2-5 3 3 + •19 •56 26 2.7 5-6 + •32 •94 6 8 - •39 1. 14 H-3 - •65 1. 91 17 1 1 . 2 3 6 + . 20 .60 4.6 6.9 + •39 1. 14 4 2 - •24 .70 9.6 - •55 1 .62 2.7 2 8 + .16 •47 27 11. 5-6 + •32 •94 8 4 - .48 1. 41 8-3 - •47 1 .40 18 9-8 4 6 + .26 •76 2.8 6.0 + •34 1 .00 10 4 - •59 1.74 4.2 - .24 •7i 20 2.7 1 7 + . 10 •29 30 2.5 3-4 + 19 •57 8 1 - .46 1-35 10.5 - .60 1 .76 21 10.6 2 4 + .14 •41 4-7 6-3 + •36 1 .02 7 1 - .40 1. 18 30 - 17 .48 3-2 5 3 + •30 .88 3i 11. 4 2.4 + .14 .41 5 4 - •3i •9i 8.9 - 51 1.50 22 12. 1 7 7 + •44 1 . 29 3-7 30 + •17 .48 6 6 - .38 1 . 11 7-7 - •44 1.29 4.0 1 8 + . 10 •3i Nov. 1 30 8-7 + •50 1 .46 6 8 - •39 1. 14 2.8 - .16 •47 23 2.7 5 9 + •34 1 .00 5-6 6-7 + .38 1. 12 1 8 - . 10 •31 6.0 - •34 1. 00 24 2-5 3 7 + . 21 .61 2 2-4 6.7 + .38 1 . 12 3 - 17 .48 9-5 - 54 1. 59 5-3 1 8 + . 10 •31 3 2.7 1.8 + . 10 •31 2 3 - •13 .38 30 - • 1 1 • 50 27 3-4 5 3 + •30 .88 50 6.0 + •34 1. 00 6 6 - •38 1 . 11 2.4 - •14 •41 5-3 7 8 + •45 132 4 10.2 t-4 + .08 •23 2 3 - •13 •38 2.4 — .14 .41 28 10.5 1 8 + . 10 •3i 2.4 2.8 + .16 • 47 4 5 - .26 •76 2.2 - . 12 36 3-o 4 8 + •27 •79 6 30 7.2 + .41 1 . 20 5 7 - •32 .96 3.0 - •17 •50 Dec. 4 12.0 5 4 + •31 •91 7 10.6 6.2 + •35 105 4 2 - .24 .70 9-5 - •54 i-59 2-7 2 4 + .14 •4 1 2.8 5-8 + •33 •97 11 9 - .68 2.0 4-4 - •25 •73 5 11. 12 7 + •73 2.15 8 30 5-3 + •30 .88 11 9 - .68 2.00 30 - •17 •50 2-7 12 + .68 2.01 9 2-5 2.3 + •13 .38 6 9 - •39 1. 14 11 . 1 - 63 1.85 6 12.2 9 5 + 54 i-59 10 11 . 1 4-7 + •27 •79 3 - •17 • 50 8-3 - •47 1.40 2.7 2 8 + .16 •47 30 1.8 + . 10 •3i 3 6 — . 20 .60 3-6 - .20 .60 7 2.7 10 5 + .60 1.76 11 10.5 3-9 + . 22 65 4 5 - .26 .76 5-4 - •3i • 9i 8 11 . 2 9 5 + •54 i-59 2-5 4.8 + .27 •79 2 4 - •14 •4 1 8.1 - .46 i-35 3-2 7 3 ♦ •42 1.24 13 12.4 6.8 + •39 1. 14 2 3 - •13 .38 6.0 - •34 1 .00 9 10.7 6 7 + •38 1. 12 3-2 •9 + 05 .16 5 9 - •34 1 .00 5-6 - •32 • - p 2 = 2 1 cm. Date. Time. P-Pz> J. P. J -1 . n X io- s . 1906. h. m. May 21 9 30 27.4 3-6 76.4 3-2 9-7 12 26 •7 2. 1 2.8 6.1 5 30 27 •3 30 •3 2.7 5-9 22 9 15 . 1 2.9 •7 3-1 8-5 3 00 3 3-1 .6 30 7-7 23 9 15 •3 29 •4 2.7 6.1 3 15 •5 4.4 . 1 3-7 16 24 9 30 .0 2.8 .0 2.8 6-3 4 00 . 1 3-5 75 .8 3-2 97 25 9 00 .2 4-3 .6 3-8 18 26 8 50 •4 4-9 •4 4.0 19 3 10 . 2 4-9 •3 4-2 22 27 10 25 . 1 4-2 .2 3-6 15 4 25 . 1 4-9 .0 4.2 22 28 9 5 27 •3 5-2 75 .0 4.2 22 3 45 . 1 5-4 74 •9 4.6 30 29 11 00 . 1 4-9 75 . 1 4.2 22 3 50 . 1 4-9 .2 4-3 25 30 8 50 . 1 2.9 •9 2.7 6 3 10 •3 3-9 .8 3-3 11 3i 9 30 3-o •9 29 7 2 45 4-5 6 4-3 25 June 1 9 30 2 3-9 7 34 12 2 45 26 9 3 9 3-9 17 2 9 5 27 3-7 6 3-5 13 2 45 1 3-7 6 3-3 11 3 10 00 3 4.9 9 4-4 26 4 30 2 4.2 9 3-8 18 4 8 30 27 1 3-i 76 4 30 8 3 35 3 4-3 2 4.0 19 5 9 20 1 3-9 75 9 3-7 16 3 20 1 4-2 6 3-8 18 6 10 6 1 50 1 4-3 24 3 20 2 59 74 8 4-9 37 7 9 10 3 4-9 75 6 4.2 22 3 45 3 4.8 9 4-3 25 8 9 25 1 3-6 76 2 3-6 15 2 45 4 4.6 4.0 19 9 9 10 3 5-o 75- 5 4-3 25 2 45 26! 9 3-5 3-i 8 8 15 27. 3 5-1 1 4.1 20 9 20 3 5-i 3 4.2 22 10 9 55 2 5-1 2 4-3 25 3 15 3 5-6 4.6 30 6 15 2 5-1 1 4-3 25 11 8 40 2 4-9 5 43 25 2 55 1 4.2 6 3-8 18 TIME VARIATIONS OF NUCLEI. 157 Table 54, continued. — Colloidal nucleation (s) of dust-free air near the fog limit in their time variations, observed two or more times daily. dp*>* 27 cm. (observed at fog chamber, isothermally) =*p- p z . Computed dp = p - p 2 = 2 1 cm. Date. Time. P-P*. s. P. J 1 . n X io-». 1906. h. m. June 1 1 6 5 O.I 3-6 0.8 3-3 11 12 8 45 27.1 3 76 3 3- 8 12 00 .2 3 4 3 3- 3 11 3 00 .0 2 8 2 2. 9 7 13 9 00 . 1 2 5 7 2 7 6 11 40 . 1 2 5 7 2 7 6 2 35 .2 3 6 3 8 6 5 .2 3 5 5 3 5 13 H 9 35 .2 3 9 2 3 7 16 12 20 .0 2 9 2 9 7 3 30 .2 4 75 8 3 6 15 6 10 . 1 3 7 7 3 4 12 15 9 20 . 1 4 2 7 3 9 18 12 00 .2 4 2 7 3 7 16 2 40 . 1 4 5 7 4 2 22 6 00 . 1 3 6 8 3 3 11 16 9 50 ■4 4 7 9 4 1 21 3 45 .2 4 1 3 7 16 17 10 00 . 1 4 6 3 6 15 12 00 •3 5 3 6 4 6 30 3 25 •3 5 1 6 4 5 28 6 10 26.8 2 8 7 3 8 18 9 15 27.1 3 76 2 9 7 12 35 .0 3 75 9 2 9 7 3 00 .2 3 9 9 3 5 13 6 00 .2 3 8 8 3 4 12 19 11 00 .1 3 5 6 3 1 8 12 55 .2 4 4 6 3 8 18 3 40 .2 4 6 6 4 19 20 5 35 . 1 4 1 7 3 8 18 21 10 15 • 3 4 8 3 .6 14 1 00 .1 3 8 7 3 •5 13 3 10 . 1 3 7 6 3 3 11 6 20 . 1 3 8 3 •4 12 22 9 15 . 1 4 4 •3 3 .8 18 12 10 . 1 4 2 •3 3 .6 15 2 50 .2 5 5 .2 4 •7 33 6 10 .2 5 3 4 •5 28 made at about 9 a. m. and 3 p. m., as near the time of the Wood and Campbell maxima and minima as my duties permitted, on the successive days and hours given by the abscissas. 99. Deductions. — Figure 64 shows in the first place that, in general, minima and maxima of nucleation would have to appear at about the time at which Wood and Campbell observed maxima and minima, i58 VAPOR NUCLEI AND IONS. an explanation on similar lines to the respectively ; or that an inversion of Wood and Campbell's results is a question, since there is usually incremented nu- cleation in the afternoon as compared with the morning. This, however, may be explained, if the ions are large, even in comparison with the larger gradations of col- loidal nuclei. Fewer of these will therefore be captured in proportion as the ionization is larger. Hence the figure shows at a an apparent corroboration of Wood and Campbell's results; at e an omission or in- version of the periods. But the e's are much fewer in number, and in comparison with the am- plitude of the a 's, the e 's are frequently neutral. In the second place, the high nucleation dur- ing the period of rain is noteworthy. Here, then, few ions were present. As there is a modification of the atmospheric po- tential gradient during this time , one might favor ideas suggested by Richardson.* 100. Effect of the barometer. — In the further development of the investigation on the time variations of the efficient colloidal nucleation in filtered air, the results are of the same character as those already dis- * Nature, lxxiii, p. 607, 1906. TIME VARIATIONS OF NUCLEI. 1 59 cussed; but the dependence of the nucleation on the fluctuations of the barometer now shows itself even more obtrusively than before. The minima of atmospheric pressure coincide with maxima of colloidal nucleation, and therefore (by inference but not necessarily) with minima of ionization of the dust-free air, both in the daily and in the weekly periods of observation. Maximum pressure, therefore, would correspond to maximum ionization as if the radiant energy originated in the com- pression of the atmosphere, or were dependent on the mass of the atmos- phere bearing on a given place. This would, if finally substantiated, be an important inference, but no more so than the more direct correlative result that minimum pressure and maximum colloidal nucleation of dust- free air go together. At the same time, since the change of absolute temperature (r) due to a sudden expansion equivalent to a drop (dp) at a barometric pressure (p) and vapor pressure (7:) may be written the correction for the changes of the barometer are in the same sense as the observed changes in nucleation. These corrections are found by varying the numerator of dpj(p-Tt) and observing the effect on the angular diameter of the corona. While I can see no room for error, it must nevertheless be acknowledged that the present method of small exhaustion, though possibly more sensitive, is not as straightforward as the method mentioned in my address, where no variation could be detected, the terminal corona remaining unchanged. At the present stage of investigation, therefore, the need of any external radiation has ceased to be obvious; and the results, if they exceed the barometer correction, are most directly referable to changes of pressure within the gas, the number of colloidal nuclei being greatest when the pressure is least. 101. Corrections. — In table 54 and the following table 55, p is the barometric pressure ; p 3 the final pressure when fog chamber and vacuum chamber are in communication; p 2 the (computed) pressure which would be observed in the fog chamber if it could be instantly separated from the vacuum chamber after exhaustion. Hence the true drop in pressure is dp =p - p 2 while dp = p - p 3 is observed. The initial pressure (p) differs from the atmospheric pressure by a few millimeters, due to the leakage of the 4-inch stopcock from fog chamber to vacuum chamber. As it is impossible to retain the same value of p and dp throughout, a correction must be applied, seeing that the adiabatic cooling is given by t 2 /t 1 = (1 -dp/(p- 7r)) ( *~ c)/ *, where n is the vapor pressure of water. By varying l6o VAPOR NUCLEI AND IONS. dp from 26 to 28 cm. and observing the changes of s (angular coronal diameter) when p is constant, a table was investigated from which reductions could at once be made to dp = 27 cm. and £ = 76 cm.; for instance, dp -26.8 cm. ds- = +0.34 P = 75.8 ds- = -O.I I 26.9 + .17 75-9 - .06 27.0 + .00 76.0 + .00 27.1 - .15 76.1 + .06 27.2 - .30 76.2 + .11 The value of s corrected for both pressures is given in this table. In the same way the case of apparatus II (to be described presently) was treated, the reductions being much larger here. Thus the data (barometer, 75.5; temperature, 21.4 C.) dp 24.9 J-1.7 dp =25.6 s -3.1 25.6 3.6 26.2 5.2 264 5.6 25.6 2.9 24.8 1 wo were consecutively observed, showing that the coronas after long waiting are much larger than when obtained in succession. The efficient nuclea- tion of the fog chamber increases in the lapse of time, probably from the evanescence of water nuclei associated even with small coronas. From these results corrections of the form dp 25.5 ds=* +0.27 £=75.9 ds — 0.08 25.6 + .00 76.0 + .00 25.7 -f .27 76.1 + .08 were made out. These examples show how critically important these corrections are, and how difficult it will be to decide whether anything more than the barometer fluctuation is being observed. 102. Further data. — It follows from section 101 that there would be a chance for error in the unavoidable leakage of the stopcock, and the best method of coming to a decision would consist in the installation of a second fog chamber, side by side with the other, but containing a smaller exhaust pipe (2 inches in diameter, 18 inches long) and therefore a more perfect stopcock. Fig. 65 shows two fog chambers (F and F f ) and appurtenances in place, each with an independent vacuum chamber (V and V) and goniometer. Other parts will be easily recognized. Table 55 and chart 66 contain the new data, the old fog chamber (4-inch cock) being marked I, the new fog chamber (2-inch cock), marked II. The latter was first tested at different values of dp, and dp = 25. 6 TIME VARIATIONS OF NUCLEI. 161 NMIHNN^d fH IPS II 13 > 1 i ' ~ ■y L I i l i -1 Dfli ir w ti.i j r i ;' i i ill III'' JL 1 ' pi IF; Or ii 1 MB • ■ "Mi t;'/ c Fig. 65.— Independent fog chambers with 2-inch and 4-inch efflux side by side. Vacuum chambers, V and V ; fog chambers, F and F'. finally selected. At the outset the values of s may be uncertain, because of internal sources of nucleation and similar difficulties. Four observations were usually made daily. Data corresponding to I and II are distinguished by subscripts, and those of II were usually observed about five minutes after the time given for I. One may note the much greater efficiency of the apparatus with 2-inch pipes as com- pared with the other with 4-inch pipes, the reason for which is not easily seen. As a whole the observations for apparatus I during the end of June and the beginning of July run parallel to the barometric decrement, and the astonishingly high values on June 29 and 30 occur during a period Table 55.— Colloidal nucleations near the fog limit in their time variations. Fog cham- ber I with 4-inch exhaust pipes; 3p=p- £3=27 cm.; II with 2-inch exhaust pipes; dp =p - p s - 25.6 cm.; p - p 2 = 21 cm. and 20 cm., respectively. Date. Time. *Pi-lhh *v P «V n X 10- 3 8P*-P-P* •y 2 . «v n 2 X 10- 8 . h. m. June 23 9 15 27.1 4-3 75 5 3.8 18 '27.6 7-3 6.4 75 12 35 . 1 4 5 6 4 1 20 4 7 7 7-4 86 3 15 . 2 4 6 6 4 1 20 4 7 3 7.0 82 5 50 .0 3 6 6 3 4 12 3 7 1 7.0 82 24 10 20 . 1 3 8 s 3 5 13 2 7 5 7-9 93 12 50 .0 4 1 7 3 9 18 2 26 5 5 2 5-o 3« 4 15 . 1 3 9 7 3 6 15 6 5 1 4.6 30 6 20 . 1 4 3 7 4 O 19 5 5 5 5-3 45 25 9 10 .2 5 1 7 4 6 30 5 6 5-8 57 12 10 . 1 4.0 s 3 7 16 •5 4 9 4-7 32 Early results. * Results in succession. l62 VAPOR NUCLEI AND IONS. Table 55. — Continued. Date. Time. tpi-P-h j. A- /«. rcx io -3 dp 2 =p-p 3 *,. •V n 2 x IO~ 8 . h. w. June 25 3 35 27.1 4.2 75-8 3-9 18 26.5 5-o 4.8 34 6 10 . 1 3 7 2 7 6 •5 5 3 5 1 40 26 9 30 27.1 4 76 3 9 17 12 20 . 1 4 1 75 9 3 9 17 26.5 4 9 4 8 34 12 35 9 25.6 2 1 2 2 2 40 27. 2 5 1 9 3 7 io .6 2 6 2 5 4 5 30 .2 4 7 9 4 3 25 .6 2 6 2 5 4 27 9 10 . 2 4 9 B 4 5 28 .6 2 7 2 5 4 12 25 . 1 4 9 7 4 6 32 .6 2 7 2 5 4 3 00 . 1 4 1 6 3 7 16 .6 3 5 3 2 9 6 5 .2 4 8 6 4 2 22 •5 3 5 3 4 11 28 9 5 •3 5 7 9 5 2 43 .6 3 3 7 12 5 . 1 4 9 9 4 7 33 •5 3 3 2 9 3 30 .2 5 1 8 4 7 33 •5 2 8 2 6 5 29 9 00 . 1 5 3 4 5 28 •7 3 7 2 9 7 12 30 ■5 7 2 5 8 59 .6 3 8 3 3 10 12 40 .2 5 1 2 4 3 25 2 55 .2 5 4 2 4 6 30 25. 6 3 7 3 2 9 6 00 . 2 9 2 4 2 22 * .6 3 8 3 2 9 30 9 25 • 5 7 2 5 8 59 •7 4 4 3 6 14 12 30 • 1 3 3 1 4 6 30 .6 3 2 2 6 5 3 5 . 1 5 3 2 4 7 33 .6 4 3 3 10 5 50 . 1 5 3 3 4 7 33 .6 4 3 3 7 16 July 1 10 00 27.2 5 1 75 6 4 6 30 25.6 3 3 3 7 12 10 . 1 4 5 6 4 1 20 .6 3 2 2 9 7 1 45 . 1 4 1 6 3 7 43 •7 3 6 2 9 7 6 25 .2 5 2 6 4 7 33 •5 2 9 2 9 7 2 9 15 •4 S 7 6 4 9 37 .6 3 6 3 3 10 12 30 . 1 4 6 6 4 2 22 .6 3 6 3 3 10 2 20 •3 5 3 5 4 8 36 .6 4 2 3 8 17 6 00 . 2 5 5 5 4 9 37 .6 3 7 3 3 10 3 9 40 •3 5 6 5 4 9 37 .6 3 8 3 4 1 1 12 30 . 2 5 3 5 4 7 33 .6 3 5 3 1 9 2 00 •3 5 3 4 6 30 .6 3 6 3 2 9 6 50 . 2 5 3 5 4 6 30 .6 3 9 3 5 13 4 9 15 . 2 5 3 2 4 5 28 .6 4 1 3 5 13 12 00 . 1 5 4 4 26 •7 4 4 3 5 13 4 50 . 2 5 2 3 4 5 28 .6 4 3 3 7 16 6 5 •3 5 2 5 4 5 28 .6 3 5 3 1 8 5 8 55 . 2 4 76 5 3 9 18 .6 2 8 3 2 9 12 10 •4 4 7 5 4 4 27 2 45 .2 3 7 5 3 7 16 .6 2 I 3 6 14 6 20 . 2 3 3 7 3 4 12 .6 1 9 2 4 4 6 9 50 27. 2 3 7 77 1 4 19 256 1 9 2 8 6 12 35 . 1 2 7 3 1 9 .6 1 8 2 6 5 12 45 _ 2 2 9 3 1 8 3 15 t 2 3 3 76 9 3 5 13 25.6 1 9 2 6 5 6 10 •3 3 8 9 3 9 19 .6 1 5 2 2 3 7 9 20 . 2 3 6 7 3 6 15 .6 2 5 3 1 8 2 25 _ 2 4 2 4 2 22 .0 1 3 3 6 H 7 45 . 2 3 7 76 4 3 5 14 .6 3 S 3 8 17 8 8 50 . 2 4 2 2 3 9 18 •4 2 8 3 5 13 9 00 . 2 3 8 .2 3 6 l 5 2 55 . 2 4 3 . 1 4 20 !6 2 7 2 8 6 5 40 •3 4.9 .0 4-4 27 .6 3-o 3-o 7 TIME VARIATIONS OF NUCLEI. >63 Table 55.— Continued. Date. Time. dpx-p-p* jfc. P- #v »X lO~ 3 dp 2 = M>3 J 2 . j' 2 . n 2 x IO - '. h. m. July 9 9 00 27. 2 5-2 75-8 4.8 35 25 4 3 3 3-5 13 2 30 .2 5 1 7 4 6 30 6 4.1 3 7 16 5 50 .2 4 8 6 4 3 25 6 4>l 3 8 17 10 9 05 . 2 5 2 6 4 7 33 7 4.0 3 4 11 2 45 •4 5 7 5 4 8 35 6 4.0 3 6 14 5 45 . 2 5 4 5 29 4 2.8 3 7 11 9 05 . 2 4 6 76 2 4 4 27 5 2.6 3 O 7 2 15 . 2 3 9 2 3 7 16 5 20 . 2 3-9 . 2 3-7 16 June 23 Fig. 66.— Data of fig. 64 continued for apparatus I (fog chamber with 4-inch pipes) and apparatus II (fog chamber with 2-inch pipes), adjusted for different exhaus- tions {dp =27.3 and 25.6 cm., respectively). Table 55. of exceptionally low barometer. The daily details of the barometer arc not as a rule reproduced. In case of apparatus II, if the initial observations ^ = 27.4 be over- looked, the data for dp = 26.5 are not unlike magnified parts of the corresponding curve for apparatus I. The chief observations for appa- ratus II in June (^ = 25.6 or computed dp =20 just above the fog 164 VAPOR NUCLEI AND IONS. limit) do not in their details, it is true, agree with the curves for appa- ratus I, except to some degree on June 29 and 30. As a whole, how ever, they also strikingly follow the march of the decrement of the barometer. One may note, moreover, that it is not necessary that the curves for apparatus I and II should quite agree. If, for instance, the preponderating nuclei in the former case are colloidal, and in the latter case ions, one should expect an inverse curve; for the ions in I would decrease the number of efficient nuclei, whereas in II they would increase their number. Decision must be deferred for further observation. In July the two fog chambers again fail to agree in their daily fluctua- tions, but both nevertheless follow the barometer closely, in their broader variations. The different sizes of coronas imply different ratios of ions and colloidal nuclei, and hence detailed agreement should not be antici- pated. On July 12 the fog chambers were subjected to modifications and it was therefore concluded to terminate the preliminary series of observations at that point, reserving further record and comment for a subsequent report. 103. Conclusion. — It appears, therefore, at the present stage of progress of the investigation on the time variation of the number of larger colloidal nuclei in dust-free wet air, that the meaning of the above results is not quite clear. The probable occurrence of an effect due to anything like cosmical radiation may, however, be regarded as excluded. The nuclea- tion curves vary in their broad contours with the barometer, being a maximum when the atmospheric pressure is least. Nevertheless the two fog chambers do not, in their detailed variation, show appreciable accordance; rather the reverse of this, and yet such a result may be due to differences in the ratios of nuclei and ions entrapped in the differently adjusted chambers. No. II in the above work (Chapter II, section 53) shows apparent excess in the region of ions as compared with No. I. As a whole the data for the time variation can not be regarded as quite trustworthy, since the correction to be added for variation of baro- metric pressure and the drop of pressure are in the same sense as the residual effect observed. If some obscure factor in the make-up of these corrections has been overlooked, the results may possibly be attributed to it. From another point of view adequate explanation is now at hand, why in the above investigations it was found impossible to secure rea- sonably coincident results in the case of those groups of nuclei which lie in the region of the larger colloidal nuclei and of the ions ; and it is chiefly for this reason that the results of this chapter are included in the present report. p THIS BOOK IS DUE ON THE LAST DAVo STAMPED BELOW RENEWED BOOKS ARE SUBJECT TO IMME $ &r/7*i*&M( 'Olr*/£3& N° 474295 $ts Baru8 ' c * Condensation of vapor as induced by nuclei and ions -^PHYSICAL QC918 B33 ^ LIBRARY UNIVERSITY OF CALIFORNIA DAVIS ills Barus, C. Condensation of vapo|r as induced by nuclei 3®$