UNIVERSITY OF CALIFORNIA COLLEGE OF AGRICULTURE AGRICULTURAL EXPERIMENT STATION BERKELEY, CALIFORNIA PLOT TESTS WITH CHEMICAL SOIL STERILANTS IN CALIFORNIA A. S. CRAFTS, H. D. BRUCE, and R. N. RAYNOR BULLETIN 648 MARCH, 1941 UNIVERSITY OF CALIFORNIA BERKELEY, CALIFORNIA CONTENTS PAGE Introduction 3 Materials and methods 5 Results 7 Discussion of results 11 Annual vegetation 11 Perennial vegetation 15 Colloidal adsorption 16 Species tolerance 18 Mixed chemicals 18 Arsenic trioxide and sodium chlorate 18 Sodium arsenite and sodium chlorate 19 Borax and chlorate 19 Colemanite and chlorate 19 Arsenic trioxide and borax 19 Practical uses of chemical treatment 19 Summary 23 Literature cited 25 PLOT TESTS WITH CHEMICAL SOIL STERILANTS IN CALIFORNIA 1 A. S. CRAFTS, 2 H. D. BRUCE, 3 and R. N. RAYNOR* INTRODUCTION Chemical weed control is becoming an accepted practice in California (figs. 1 and 2). In agricultural regions the use of chemicals should gen- erally be restricted to situations under which tillage and cropping methods are not effective or economical. Since, however, weeds may be hosts for insects and organisms causing plant diseases and are potential fire hazards, there is often need for complete sterilization of soils in waste or unused areas. Chemical treatments for weed control include the use of (1 ) selective and nonselective contact sprays for annual weeds, (2) translocation sprays for killing deep-rooted perennial plants with- out injuring the soil, (3) temporary soil sterilants for controlling peren- nials on agricultural land, and (4) permanent soil sterilants for clearing firebreaks, ditchbanks, fence lines, and roadsides of both annual and perennial vegetation. It is with these last two treatments that this paper is concerned. As shown by previous greenhouse tests {2-9, 13) , 5 the effect on vegeta- tion of chemical treatment through the soil depends on the following factors: (1) inherent toxicity of the chemical; (2) adsorption of the chemical by the soil; (3) decomposition or chemical change tending to reduce toxicity; (4) leaching of the chemical; (5) soil fertility or salt concentration (intimately related to chlorate toxicity) ; and (6) spe- cies tolerance. In the open field, net toxicity is influenced not only by these but by factors not usually involved in greenhouse experiments — for example, root distribution, wind and water erosion, treading by animals, burrow- ing by rodents, nonuniform distribution of the chemical, and variability of plants and plant species. Because such uncontrolled variables greatly affect actual results in the field, greenhouse studies must be supple- mented with plot tests before any particular treatment is put to prac- tical use. 1 Received for publication October 16, 1940. 2 Associate Professor of Botany and Associate Botanist in the California Agricul- tural Experiment Station. 3 Chemist, California Forest and Range Experiment Station, Forest Service, United States Department of Agriculture. 4 Associate in the California Agricultural Experiment Station. 5 Italic numbers in parentheses refer to "Literature Cited" at the end of this bul- letin. £ 3 -j University of California — Experiment Station Fig. 1. — Successful soil sterilization on the banks of a concrete-lined irrigation ditch. (From Biennial Report, 1932-1934.) Fig. 2. — In the absence of chemical or mechanical treatment, ditch-bank weeds interfere with the flow of water and drop millions of seeds into the water for distri- bution on fields. (From Biennial Report, 1932-1934.) Bul. 648] Plot Tests with Chemical Soil Sterilants 5 To develop practical field methods of weed control, over 1,200 plot tests with chemicals have been made in California during the past eight years. Through cooperation between the Botany Division of the College of Agriculture, University of California, the California Forest and Range Experiment Station of the United States Forest Service, and the Klamath Weed Control Project of the California Agricultural Experi- ment Station, herbicidal chemicals have been tested in the agricultural areas of the central valleys, in the forests of the Sierra Nevada, and in the foothill ranges of Humboldt County. Their action has been observed on 13 soil types. The principal chemicals used were sodium arsenite, arsenic trioxide, borax, colemanite, 8 sodium chlorate, and ammonium thiocyanate. This paper summarizes the combined data from the plot tests by the three cooperating agencies to indicate in a general way the responses of weeds to these chemicals and the dosages in pounds per square rod found necessary for soil sterilization in practical weed control. MATERIALS AND METHODS The plots in nearly every instance were one square rod in area, staked at the corners and marked for future identification. Most of the plots were not fenced, although some in the forest and range regions were within enclosures. In some instances the existing vegetation was burned off before treatment ; some of the plots in Humboldt County were cleared by hoeing. Usually, however, the chemicals were applied without pre- liminary preparation. Technical grades of the various chemicals were used. Sodium arsenite was made, in most cases, by mixing arsenic trioxide, sodium hydroxide, and water in the proportions of 4 : 1 : 3 by weight and diluting to a con- venient concentration. Arsenic trioxide of 95-98 per cent purity was used in the dry form just as purchased from the smelters. The borax was a technical granulated grade assaying about 36 per cent B 2 3 , and the colemanite was a finely ground sodium-calcium borate ore of 28 to 30 per cent B 2 3 . In most instances both borax and colemanite were used dry. Technical sodium chlorate was used both dry and dissolved. Am- monium thiocyanate, as manufactured in trial lots for weed-control pur- poses, was used in solution. The dry chemicals were applied by hand, the solutions by spraying, with as uniform distribution as possible. In nearly every experiment, each chemical was applied to a series of plots in graded amounts designed to provide a range from a low to a high degree of control. Figures 3 and 4 illustrate typical plots several years 8 The boron ore used in this work was not pure colemanite but a mixture of cole- manite and ulexite. 6 University of California — Experiment Station after treatment. The plots were periodically inspected, and visual esti- mates were made of the proportion of vegetation on each plot as com- pared with adjacent untreated areas. Separate estimates were made for Fig. 3. — Sterilization plot four years after treatment with ar- senic trioxide (4 pounds per square rod). The heavy line demarks the plot boundary. annual and perennial species. Table 1 exemplifies the recorded data from 42 of these plot experiments. Table 2 presents information on the 13 soils. The textural grades and clay contents 7 were determined on samples taken from the top 6 inches after grass and debris had been removed. 7 The writers are indebted to the Bureau of Chemistry and Soils, United States Department of Agriculture, for the soil analyses. Bul. 648] p L 0T Tests with Chemical Soil Sterilants 7 RESULTS The data are graphically summarized in figures 5 to 10. "Percentage Kill" in these figures means the amount of vegetation that the chemical prevented from growing, expressed as a percentage of the normal growth on adjacent untreated check areas, determined by visual estimate. Dos- Fig. 4. — Sterilization plots seven years after treatment with arsenic trioxide. The heavy lines demark the borders of plot areas in which applications were 8, 12, and 16 pounds per square rod. Vegetation has returned to the right hand portions of the plots where road maintenance machinery has disturbed the soil. ages in all cases are expressed in pounds per square rod. In preparing these graphs, observations on percentage kill from parallel series of plots were averaged and graphically smoothed, successively on dose, on year, and, where soil texture was important, on clay content. The relations shown in the graphs of figures 5 to 10 result from averag- ing many individual readings, among which there was considerable variation. The herbicidal effect in any one instance may therefore differ appreciably from that shown on the curve. To illustrate, table 3 records the standard errors of estimate in percentage kill of annuals by arsenic trioxide. That appreciable deviations occur may be expected in view of University of California — Experiment Station table l Data Typical of Observations from Plot Tests with Soil Sterilants (Davis, California; October 21-24, 1931) Pounds chemical applied per square rod Percentage kill, as of dates given Plot no. April 29, 1932 March 30, 1933 April 21, 1934 Annuals Perennials Annuals Perennials Annuals Perennials Arsenic trioxide, AS2O3 1 1 2 10 10 2 3 4 10 50 50 4 8 50 80 75 5 12 16 60 80 85 90 85 95 6 AS2O3, as sodium acid arsenite in solution; As203:NaOH = 4:l 7 1 8 2 4 8 50 80 95 10 40 97 20 25 9 10 11 12 100 5 98 90 12 16 99 10 99 5 95 Sodium chlorate, NaC103, in soluti 19 1 2 4 8 12 16 25 40 60 80 90 90 50 80 90 100 100 100 25 60 75 25 70 80 90 95 95 10 25 20 21 22 25 75 23 85 24 90 Ammonium thiocyanate, NH4CNS, in solution 37 1 2 4 8 12 16 10 10 25 50 50 75 5 15 25 50 15 25 15 38 39 40 41 42 Bul. 648] Plot Tests with Chemical Soil Sterilants 9 the adventitious factors mentioned above and also the disregard of such variables as rainfall and species tolerance. Annual species are, in general, more nearly alike in their responses to toxic chemicals than are perennials, probably because they have rela- tively shallow roots and propagate by seeding. They can therefore be classed together, year after year, without differentiation of species, as in the graphs of figures 5 to 9. TABLE 2 Location, Type, and Clay Content of the Soils Treated with Sterilizing Chemicals Plot location Soil name* Textural grade Per cent clay Durham Ferry road near Stockton Castle road near Stockton Dinuba Dinuba Dinuba Dinuba and Fresno Sorrento Sorrento Yolo Yolo Hugo (?)t Hugo (?)f Mariposa (?)t Diamond springs (?)t Holland Loamy fine sand Fine sandy loam Sandy loam Loamy fine sand Clay Loam Loam Clay Loam Gravelly loam Loam Loam Sandy loam 3.4 5.3 4.6 West of Turlock East of Patterson 3.5 32.4 14.2 Cache Creek canal, Davis Poultry plots, Davis 18.3 30.7 16.9 19.0 18.1 10.8 8.6 * Named by Shaw according to his classification (16) after examination of soil samples from the plots, t Four of the plots were located in areas designated as unclassified hill land. Annuals found on these plots included common foxtail (Hordeum murinum), filaree (Er 'odium species), bur clover (Medicago hispida), common chickweed (Stellaria media), miner's lettuce (Montia perfoli- ata), shepherd's purse (Capsella Bur 'sa-past oris) , common peppergrass (Lepidium nitidum), mustards (Brassica species), fiddleneck (Am- sinckia Douglasiana) , annual bluegrass (Poa annua), mallow (Malva species), yellow star thistle (Centaurea solstitialis) , soft chess (Bromus mollis), red brome (Bromus rubens), a fescue (Festuca dertonensis) , and slender oat (Avena barbata). Although perennial growth was observed on all plots for several con- secutive years, the return of perennial vegetation so largely depends on the mode of propagation of the particular species that the data are pre- sented for the first year only. The perennials observed were Klamath weed (Hypericum perforatum), bearmat (Chamaebatia foliolosa) f al- kali mallow (Sida hederacea), Bermuda grass (Cynodon D act y 'Ion) , and hoary cress (Lepidium sp.). Perennial species differ widely in their sus- ceptibility to toxic chemicals, because of inherent differences both in 10 University of California — Experiment Station 3 6 9 12 15 3 6 9 12 15 POUNDS OF CHEMICAL PER SQUARE ROD Fig. 5. — Toxic effect of several chemicals on annual vegetation, irrespective of soil characteristics. Bul. 648] Plot Tests with Chemical Soil Sterilants 100 1st Year 11 3 6 9 12 15 3 6 9 12 POUNDS OF ARSENIC TRIOXIDE PER SQUARE ROD Fig. 6. — Toxic effect of arsenic trioxide on annual vegetation in soils of various percentages of clay. their protoplasmic constitution and in the depth and distribution of their roots. Had poison oak (Rhus diversiloba) or bracken fern (Pteris aquilina), for example, been taken as indicators, the percentage-kill results would have been decidedly lower than those shown in figure 10. DISCUSSION OF RESULTS In field work with herbicides, 100 per cent kills are seldom encountered and results between 90 and 100 per cent are usually considered satisfac- tory. As shown by the shape of the typical toxicity curves of figures 5, 6, 8, 9, and 10, the last increments of chemical are the least efficient. Thus (fig. 10) over 2 pounds of chlorate was required to cause the last 5 per cent of killing, as contrasted with 65 per cent for the first 2 pounds. Annual Vegetation. — As figure 5 shows, different chemical herbicides vary widely in their toxic effects. Sodium arsenite is rapid and potent, 12 University of California — Experiment Station 100 80 60 O 40 20 FIRST YEAR Pounds As 2 Oj per Square Rod 10 15 20 25 30 100 Pounds AspO, per Square Rod Lo 10 15 20 PERCENTAGE CLAY IN SOIL Fig. 7. — Influence of the clay content of the soil on the toxic effect of arsenic trioxide on annual vegetation. losing its toxicity only slowly. Arsenic trioxide, after a delayed start, is highly effective — the most nearly permanent chemical tested against shallow-rooted plants. Sodium chlorate, in the absence of rapid excessive leaching, is effective against annuals the first year ; but the effects soon disappear and may be entirely gone by the third year. Ammonium thio- cyanate is less toxic than chlorate and even more transient. Borax and colemanite, though less toxic than chlorate, are longer lasting. 3 6 9 12 15 3 6 9 12 15 POUNDS OF CHEMICAL PER SQUARE ROD Fig. 8. — Toxic effect of sodium arsenite on annual vegetation in soils of various percentages of clay. 100 8 Percent Clay 80 60 O 40 20 4 6 8 10 12 POUNDS BORAX PER SQUARE ROD Fig. 9. — Toxic effect of borax on annual vegetation the first year after application. 14 University of California — Experiment Station CD < 100 80 60 40 20 100 80 60 40 O S 20 100 80 60 40 20 °0 Fig. 10.- Sodium Arsenite Arsenic Trioxide / Sodium Chlorate Ammonium Thiocyanate / Borax Colemanite ^^^ 3 6 9 12 15 3 6 9 12 POUNDS OF CHEMICAL PER SQUARE ROD 15 -Toxic effect of various chemicals on perennial vegetation the first year after application, irrespective of soil characteristics. Bul. 648] p L0T Tests with Chemical Soil Sterilants 15 Sodium arsenite is very soluble in water and was always applied in aqueous solution. Arsenic trioxide, on the contrary, is only slightly solu- ble and was always applied dry. This difference in solubility probably explains the superiority of sodium arsenite the first year and its gradual tendency to disappear by leaching from the top soil in subsequent years. The trioxide, being relatively insoluble, acts slowly at first, but is long present to prevent the survival of sprouting seeds. Similar differences in solubility may explain the differences in behavior between borax and colemanite. The former is more soluble, with more rapid but less per- manent action, than the latter. For treating annual vegetation the season of application is important, particularly with arsenic trioxide and sodium chlorate. Judging from TABLE 3 Standard Errors of Estimate in Percentage Kill of Annuals by Arsenic Trioxide as Compared with Observed Values Standard errors of estimate in low and high toxicity ranges 0-90 per cent kill 90-100 per cent kill Season 1 27.5 10.5 16.8 Season 2 5.9 Season 3 5.4 Season 4 1.9 experience, if arsenic trioxide is spread in early fall, the winter rains will render it more effective by the next spring than is shown in figure 5. Sodium chlorate, on the contrary, when used against annuals and shal- low-rooted perennials, should be applied not in the fall, but in the spring after ail seeds have germinated, so that subsequent showers will leach it into the zone of absorbing roots. For the chlorate curve of figure 5, data for January to April applications only were used. Perennial Vegetation. — Figure 10 shows averaged experimental re- sults on perennial vegetation. The ineffectiveness of arsenic trioxide is in marked contrast to its potency on annuals (fig. 5). Sodium arsenite also is considerably less successful than against annuals. Though the success on perennials reported here may be partially explained by leaf absorption and translocation, sodium arsenite applied during the fall or early winter to shallow-rooted Klamath weed and bearmat (the two principal perennials treated) usually causes considerable root killing as the chemical leaches into the top foot of soil. Had other common deep- rooted perennials (such as wild morning-glory, alkali mallow, hoary 16 University of California — Experiment Station cress, or poison oak) been the only perennials involved, the curves in figure 10 for sodium arsenite would have been much lower. The same is true for borax and colemanite. For poisoning perennials by soil penetration and root absorption, sodium chlorate is excellent. In interpreting figure 10, one should re- member that kill by chlorate will depend greatly upon the species. Four pounds of chlorate per square rod gives practically complete kills of Klamath weed or bearmat {12). On wild morning-glory {Convolvulus arvensis) and Kussian knapweed {Centaurea repens), 6 pounds would be needed. When used against hoary cress or alkali mallow, even more would be required; and 40 pounds might not clear a square rod infested with poison oak. The 9-pound requirement shown in figure 10 is an aver- age value for many perennials, including some of the more refractory species. Precipitation and time of application are important considerations in treating perennials through the soil. The chemical should be leached to the depth of the main root system by the time the roots are active in the spring. Success depends upon proper leaching, which requires a certain amount of rainfall after application. One pound of chlorate per square rod will usually kill Klamath weed if followed by 2 to 10 inches of rain before the spring growing season. For borax, 7 to 17 inches of rainfall appear optimum ; for colemanite, 10 to 20 inches. Three pounds of chlo- rate in the fall followed by 10 to 16 inches of rain will kill wild morning- glory but annuals will appear in the spring. This amount of chlorate applied in the spring and leached with 1 to 3 inches of rain will sterilize against nearly all annuals, but fails to injure wild morning-glory. Colloidal Adsorption. — The contrasting behaviors of arsenite, borates, and chlorate, aside from differences in inherent toxicity, depend largely upon their differential fixation by the soil {3, 7-9, 13) . Arsenic and borax are both fixed by the soil ; but, as might be expected from their chemical properties, the arsenite is very strongly held and the borate only mod- erately so. Chlorate remains quite free in the soil solution. Arsenite is strongly held against the leaching force of infiltering water ; chlorate is very subject to leaching ; borate is intermediate in behavior. The effects of colloidal adsorption of arsenic upon the toxicity of arsenic trioxide are shown in figures 6 and 7. In interpreting these curves one must consider two phenomena — adsorption and leaching. The heavier the soil, the stronger the adsorption and the lower the toxicity. The lighter the soil, the greater the leaching for a given rainfall and the lower the toxicity in the top soil. Because of these opposing factors there is a maximum in the intermediate textures. This is graphically shown Bul. 648] Plot Tests with Chemical Soil Sterilants 17 in figure 7, where percentage kill is plotted against percentage clay in the soil. Similar results were obtained for sodium arsenite except that the effectiveness is greater during the first year and lowers more rapidly with time (fig. 8). m ^mmmsr &.«?' • - ■ ■■'.,«. «\. ■> ■* v* $» % , " v v Vir - g^iaai - < ..3* ' v ** ■ > JS ^kllra • - W ^ ^- ** '< *' 1 ? - : j '<"/ :" ■*$ S^*l ^1 ?*lm?-<* ** ~ ^ ... tf.^V/'" ';' ■:*:%■-:■• ''■<. ■ ■'..' '.^:>Hyk ill? s ' ,- ^ ~ \* ... mmw^m KK ' \ * ' . -- ^ y/ Bb^^^^5^% !*<%5 :V ' .'■:,*" ' ^7 'l v .^^ -<-~ : ':' l^-'":C H§Er\ -; ■■•■' '' ^£$$&i&£j0> ..,- i ;^s^ : ; :. " " '^ ■-.... ;•. ,..' »• ! -Av' '':.,.:^' ; t*C''^* I |31yS:;::: : ;.: : C&.-\''> ■'■? :.,'. |.f ■• > ■:', : . : /:;,;.<- ■".-. _' ■ ? : ^- * "' * * i£»- " <"*,'■*'.'-."' ' : " " 4-\ ■ : '\'. V 'r'Pl '-' - sss&is*?; jsv. ;''\<; : >S>. > ' : : . ]$0& ; Kjj^^^BK^^^> * -'-- " ""■ ~-\ '(. * * i •! . ■ .'...■"'. N \ : '■ i i^^^tri "■■ ' '"\ Fig. 11. — Plot treated with 4 pounds of borax per square rod on January 6, 1934. Photographed April 26, 1935. Com- mon foxtail has survived the treatment and thrives through lack of competition and perhaps through stimulation. Borax shows a similar relation to colloid content except that the maximum toxicity seems to occur in a lighter textural grade (compare figures 6 and 9 ) . The explanation may be that many of the borax treat- ments upon which these curves are based were made in the winter months, and such early application increases the importance of leaching as a factor. 18 University of California — Experiment Station No correlation between chlorate toxicity and percentage clay was found. One might expect that the type as well as the percentage of the colloid in the soil would affect the toxicity of chemicals that tend to be adsorbed. Reasoning by analogy from the behavior of phosphates {1, 11), the kao- linitic clays should render much arsenic unavailable to plants. Accord- ing to greenhouse tests, Aiken, Sierra, and several other red soils show low arsenic toxicity to Kanota oats (9). Many such soils are character- ized by kaolinitic clay, and the low response to arsenic added to such soils is undoubtedly attributable to this fact. Species Tolerance. — In these plot tests, interesting cases of species tolerance were observed. Though arsenite solution is extremely toxic to most plants, certain species tolerate considerable quantities in the soil — notably bractscale (Atriplex bract eosa), yellow star thistle, saltgrass (Distichlis spicata), hayfield tarweed (Hemizonia fasciculata), and Turk's rug (Chorizanthe staticoides) . Seeds of these species will germi- nate and grow on soils completely sterile to many other plants. Although boron compounds will harm many vegetable and field crops, certain weeds tolerate high concentrations in the soil. Common foxtail on plots at Davis survived treatment with 16 pounds of borax or of colemanite to the square rod. Although 4 pounds per square rod elimi- nated broad-leaved plants, foxtail thrived better on plots treated at this spreading rate than on untreated areas (fig. 11). Though chlorate is very toxic to young seedlings, old established plants of certain deep-rooted perennials — notably hoary cress, salt grass, and camel thorn (Alhagi camelorum) — have survived doses as high as 8 pounds per square rod without serious injury. MIXED CHEMICALS Many hundreds of treatments with two or more chemicals were included in the plot tests. In general, herbicidal chemicals act independ- ently, and the dosage of each must be determined by the specific require- ments for its action. Where, for instance, deep-rooted perennials occur, chlorate or borax may be mixed with arsenic, the former to kill the plants already established, the latter to insure against their return for several years. Thus it has proved advantageous to use combinations of chemicals to utilize different desirable effects of each, and certain outstanding con- clusions will be described. Arsenic Trioxide and Sodium Chlorate. — Because arsenic trioxide has little effect during the year of application (fig. 5), it is often ad- vantageously mixed with chlorate, to combine its permanence with the Bul. 648] Plot Tests with Chemical Soil Sterilants 19 quick herbicidal action of the latter. This mixture may be used to kill deep-rooted perennials as well as to sterilize against annual vegetation. Since slow reaction between arsenic trioxide and sodium chlorate results in spontaneous combustion (10), mixtures of these chemicals should never be stored. Sodium Arsenite and Sodium Chlorate. — Although sodium arsenite kills annual vegetation rapidly, it does not ordinarily reach deep-rooted perennials. For this latter purpose, sodium chlorate should be added. As with arsenic trioxide, chlorate acts independently of the arsenite; each is used in the appropriate dosage to accomplish its specific pur- pose. These chemicals do not react and may be mixed in the same spray solution. Borax and Chlorate. — Where arsenic should not be used because of poison hazard to men and animals, sterilization of less permanency can be attained with borax. To many plants, however, particularly those native to arid regions, borax is not particularly toxic, and its toxicity is much greater toward seedlings than toward old established plants. To augment the initial effect, chlorate may be added to the borax. Borax alone is most successfully used against Klamath weed, bearmat, and other plants of mesophytic habit. Against the mixed annual vegetation of the central valleys and foothill regions of California, borax is insuffi- cient and should be fortified with sodium chlorate. Colemanite and Chlorate. — Colemanite, being less soluble than borax, is slower acting, but somewhat more lasting in regions of light or me- dium rainfall, though where annual precipitation exceeded 30 inches, little difference could be detected. For mixed annuals colemanite is not sufficiently toxic and may well be fortified with chlorate. Arsenic Trioxide and Borax. — Under certain conditions, chlorate is inadvisable because of its tendency to incite spontaneous ignition. In such cases chlorate may be replaced with greater amounts of borax, the arsenic and borax being applied together. The former provides lasting sterilization against annual vegetation, the latter rapidly kills suscepti- ble annuals and perennials. In the moist soil there is evidence of some chemical interaction between these chemicals. PRACTICAL USES OF CHEMICAL TREATMENT Herbicidal chemicals find practical use in eradicating noxious peren- nials on pastures, ranges, and cultivated fields and in sterilizing soils against weed growth on firebreaks (fig. 12); along fences (fig. 13), ditches (fig. 14), or roadsides; and around campgrounds, electric sub- stations, irrigation structures (fig. 15), or telephone poles (fig. 16). 20 University of California — Experiment Station Figures 5 to 10 indicate the chemicals and dosages best suited to par- ticular cases. The generalized interpretations of these graphs, tempered by the judgment of experience, should be a practical guide to the re- sponses of plants to these toxicants. Fig. 12. — Sterilized firebreak through bearmat in yellow-pine area of the Sierra Nevada. Treatment: equal parts of arsenic trioxide and sodium chlorate, applied dry at the rate of 8 pounds per square rod. The longevity of herbicidal effectiveness will depend upon soil type, rainfall, weed species, and erosion, as well as upon the chemical and the dosage. Light annual applications of chemicals are wasteful of labor, and very heavy treatments designed to last for many years may be wasteful of materials because of leaching and erosion. A practical com- Bul. 648] Plot Tests with Chemical Soil Sterilants 21 Fig. 13. — A 1-square-rod plot on a fenceline sterilized with sodium arsenite. After two rainy seasons this plot is practically free of weeds. ., . - ••■■"••S.V.-: ■•.,-...•• ■?',?:;.&;■: i%:.;*;.:<-ji w ,i. ; ^::;:',r:-,-. Fig. 14. — The main canal between lifts of the West Stanislaus Irrigation District. The banks were sterilized with 5 pounds per square rod of sodium arsenite to keep weed seeds from the water. promise would be to use the various chemicals initially at intermediate dosages, according to the relative labor and material costs, making sup- plementary applications when necessary. Since arsenic compounds are poisonous, certain precautions must be exercised in their use. They are not only deadly when taken internally, 22 University of California — Experiment Station but in the pores and beneath the fingernails they produce local irritation and soreness. Animals must not be allowed to graze upon forage dusted or sprayed with arsenicals until the chemical has been washed off by rain. Once in the soil, this poison is no longer accessible to animals. Fig. 15. — Upper: a weedy ditch. Lower: structures on the main canal of the West Stanislaus Irrigation District. Sterilization has eliminated undesirable weed growth, as shown in the foreground. (From Ext. Cir. 97.) Chlorates are not extremely poisonous to animals like the arsenicals, but they have been known to incite spontaneous ignition. As a conse- quence, chlorate should not be used in very hot, dry weather and care should be taken not to spray chlorate solutions onto finely divided acid organic matter such as decaying leaf mold. Clothing moistened with Bul. 648] Plot Tests with Chemical Soil Sterilants 23 chlorate and dried is very inflammable and may be readily ignited by a flame or friction spark. Clothing, shoes, and gloves shonld be kept free from chlorate, but if accidentally saturated they should be kept moist until washed. There should be no smoking where chlorates are in storage or where they have recently been applied. SUMMARY Observations on soil sterilization from over 1,200 plots on 13 different soils have been analyzed to show typical herbicidal behavior of sodium Fig. 16. — Arsenic sterilization against annual weeds around telephone post as a fire-prevention measure. arsenite, arsenic trioxide, sodium chlorate, ammonium thiocyanate, borax, and colemanite toward annuals and certain perennials (figs. 5-10). Sodium arsenite solution is rapid in its toxic action and decreases in effectiveness only slowly over a period of years. An application equiva- lent to 5 pounds of As 2 3 per square rod will kill, on an average, 95 per cent of annual growth the first year and 80 per cent the fifth year. On perennials sodium arsenite is less effective, but applications equivalent to 10 pounds or more of As 2 3 per square rod will usually kill nearly all shallow-rooted species. Arsenic trioxide applied dry is not highly effective the first year, but thereafter it is even more potent against annuals than is sodium arsenite. Thus, without distinction as to soil and species, 5 pounds of arsenic triox- 24 University of California — Experiment Station ide uniformly distributed over 1 square rod will normally kill about 95 per cent of annual vegetation from the second to the sixth year. On most perennials arsenic trioxide is practically without effect. Sodium chlorate, acting through the soil, is very poisonous to both annuals and perennials. It is very soluble and is readily leached. For highest success consideration should be given to the rainfall between the time of application of chlorate and the season of active root absorp- tion. The depth of penetration of the chemical into the soil should coin- cide with the depth of the main root system at the time of maximum root activity in the spring. Ammonium thiocyanate, like sodium chlorate, is very soluble ; but it seems to possess less inherent toxicity than sodium chlorate. It also de- composes rapidly in the soil (14) and has occasionally been observed to stimulate weed growth. Borax and colemanite have somewhat more lasting effects than chlo- rate but less than arsenic. Though they do not possess the high inherent toxicity of these other herbicides, against Klamath weed and bearmat they have proved very successful. Being nonpoisonous to livestock, they can be used without hazard on ranges. Sodium chlorate added to arsenic trioxide or to sodium arsenite com- bines its rapid action with the lasting properties of the arsenicals. Since chlorate acts on deep-rooted perennials, the mixture offers an effective treatment where both deep- and shallow-rooted plants abound. Observations on these test plots substantiate greenhouse and labora- tory results in showing that the following variables pertain to chemical weed control : (1) inherent toxicity of the chemical ; (2) colloidal adsorp- tion of the chemical by the soil ; (3) decomposition of the chemical ; (4) removal by leaching; (5) salt content of the soil; and (6) tolerance or physiological resistance of various weed species to toxic chemicals. The results of these experiments provide information of value for practical weed control. For large infestations, cropping and cultural methods are available ; for small areas, initial infestations, and locations difficult of access, chemical treatments are usually preferable. The in- formation here offered should aid in control of vegetation for fire pre- vention and in the elimination of noxious infestations from fertile agricultural lands. Bul. 648] Plot Tests with Chemical Soil Sterilants 25 LITERATURE CITED 1. Bukj>, John S., and H. F. Murphy. 1939. The use of chemical data in the prognosis of phosphate deficiency in soils. Hilgardia 12(5) : 32-40. 2. Crafts, A. S. 1935. Plot tests with sodium arsenite and sodium chlorate as soil sterilants in California. California Dept. Agr. Mo. Bul. 24(4, 5, 6) : 247-59. 3. Crafts, A. S. 1935. The toxicity of sodium arsenite and sodium chlorate in four California soils. Hilgardia 9(9) :459-98. 4. Crafts, A. S. 1935. Physiological problems connected with the use of sodium chlorate in weed control. Plant Physiol. 10(3) : 699-711. 5. Crafts, A. S. 1939. Toxicity studies with sodium chlorate in eighty California soils. Hil- gardia 12(3) : 229-47. 6. Crafts, A. S. 1939. The relation of nutrients to toxicity of arsenic, borax and chlorate in soils. Jour. Agr. Ees. 58(9) : 637-71. 7. Crafts, A. S., and C. W. Cleary. 1936. Toxicity of arsenic, borax, chlorate, and their combinations in three California soils. Hilgardia 10(10) : 399-413. 8. Crafts, A. S., and K. N. Baynor. 1936. The herbicidal properties of boron compounds. Hilgardia 10 (10) : 343-74. 9. Crafts, A. S., and B. S. Bosenfels. 1939. Toxicity studies with arsenic in eighty California soils. Hilgardia 12(3) : 177-200. 10. Johnson, Ethelbert. 1940. The storage of sodium chlorate and white arsenic mixture a dangerous practice. California Dept. Agr. Bul. 29(2) : 110-11. 11. Murphy, H. F. 1939. The role of kaolinite in phosphate fixation. Hilgardia 12(5) : 341-82. 12. Baynor, B. N. 1937. The chemical control of St. Johnswort. California Agr. Exp. Sta. Bul. 615:1-38. 13. Bosenfels, B. S., and A. S. Crafts. 1939. Arsenic fixation in relation to the sterilization of soils with sodium ar- senite. Hilgardia 12(3) :201-29. 14. Sandhoff, Allan G., and C. E. Skinner. 1939. The nitrification of ammonium thiocyanate (a weed eradicant) and the . effect of this compound upon the soil population. Soil Sei. 48(4) : 287-94. 15. Shaw, Chas. F. 1937. Some California soils and their relationships. Univ. of California Sylla- bus Series. Syl. J. D. Univ. of California Press 1-117.