Division of Agricultural Sciences UNIVERSITY OF CALIFORNI ui, x\' IH *"'/, \|"'/. \\i; o»/i /// \"'/ \f',/ v'h »»>!/ HERBICIDAL PROPERTIES i 0F ARSENIC m TRIOXIDE \ v '\i,"/, A. S. Crafts C. C. Buck CALIFORNIA AGRICULTURAL EXPERIMENT STATION BULLETIN 739 HERBICIDAL Pushcart applicator used for spreading ar- senic trioxide on some of the tests described in this bulletin. This particular applicator was designed in the University of California's De- partment of Agricultural Engineering at Davis. WEED CONTROL THROUGH PREVENTION can be achieved by chemical soil sterilization. EFFECTIVE, LONG-LASTING, SAFE, AND ECONOMICAL, arsenic trioxide (white arsenic) has proved to be a satisfactory means of soil steriliza- tion. On farming lands, along fence lines, ditchbanks, railroad ballast, and firebreaks, around airfields and construction or parking areas — wherever weeds are a threat to efficient operation, arsenic trioxide will destroy growth and prevent recurrence for periods of 5 to 10 years. ARSENIC TRIOXIDE HAS THESE ADVANTAGES over the soluble forms of arsenic — Not attractive to livestock, it is less hazardous. Low in solubility, it is persistent in the soil. Three forms — the original dust, a dustless powder, and pellets — make it adaptable to different kinds of terrain. It can be applied either by hand or machine depending on the terrain. This bulletin tells how 80 California soils responded to treatment in greenhouse tests, small field plots, and large-scale operations. The Authors: A. S. Crafts is Professor of Botany and Botanist in the Experiment Station, Davis. C. C. Buck is in charge, Division of Forest Fire Research, California Forest and Range Experiment Station, Forest Service, U. S. Department of Agriculture. February, 1954 PROPERTIES of ARSENIC TRIOXIDE A. S. CRAFTS C C. BUCK THE PROBLEM . . . Before arsenic trioxide could replace the more hazardous so dium arsenite as a soil sterilant, years of experimentation were needed to find out exactly how this chemical behaved in soil and how to handle it in field-scale operations. Here are the findings and recommendations concerning its use. 1 Weed control by chemical means has advanced rapidly since it was found that pesticides could be economically applied in quantity for large-scale control oper- ations. Chemical methods, long viewed as expedients, have proved practical and economical for field-scale use on mil- lions of acres in this and other countries. They have increased the possibility of waging total war on agricultural and industrial plant pests. Chemical soil sterilization is preven- tive weed control and, as in the case of medicine, prevention is better than cure. As weed control becomes more highly mechanized and as the use of chemicals increases, emphasis is bound to shift 1 Research reported in this bulletin was fi- nanced by the American Smelting and Refining Company and conducted by the Botany De- partment, University of California, Davis, Cali- fornia. Work on firebreaks was done by the California Forest and Range Experiment Sta- tion, Forest Service, U. S. Department of Agriculture. 2 Soil sterilization by means of arsenic has been described by Crafts (19356, c) ; Crafts and Cleary (1936); Crafts (19396); Crafts and Rosenfels (1939) ; Crafts, Bruce, and Ray- nor (1941); Rosenfels and Crafts (1939); Robbins, Crafts and Raynor (1942). from seasonal treatment to lasting pre- vention. In this phase arsenic steriliza- tion will occupy an important position. 2 Arsenic is still the most practical and persistent material for permanent soil sterilization, and for this purpose arsenic trioxide (white arsenic) has been found to have distinct advantages over the soluble forms. Apparently not attractive to livestock, it is less hazardous to use. It is low in solubility and is, therefore, persistent. It may be applied in the dry form either as the original dust, as a dustless powder, or in the form of pel- lets. For large-scale application the car- rier-free dry form is most economical; for treating railroad ballast the pellets have distinct advantages because they gravitate into the crevices, resisting dis- placement by wind and concentrating in the regions where seeds germinate. Arsenic trioxide, commonly known as white arsenic, is a white, finely crytal- line substance having a density of 3.75. It is produced as a by-product in the smelting of copper ores, and at times it has accumulated in huge surpluses. It is used as the primary source material in the production of sodium arsenite [3] (weed-killer), calcium arsenate dust used in cotton bollworm control, and lead arsenate used against codling moth of pome fruits. Sodium and calcium ar- senites are used as insecticides and fun- gicides, and Paris green is a mixed salt containing arsenic. There are other uses in agriculture and industry, but the above have in the past consumed the bulk of the arsenic produced in this country. With the introduction of DDT and many other new organic insecticides and the substitution of 2,4-D, borax, chlo- rates, and other chemicals for sodium arsenite as a weed-killer, it seemed pos- sible that arsenic would again be in sur- plus supply. Realizing the potential large-scale use of arsenic trioxide as a soil sterilant, scientists began work in 1949 to clarify some of the problems in- volved and to amplify our general knowl- edge of the behavior of this chemical in the soil. Arsenic is extremely toxic to plants; it kills through contact with either foli- age or roots. Applied to the tops of plants in soluble form (sodium arsenite solution), it causes rapid death of con- tacted tissues; any residue falling on the soil will cause injury to plant roots if carried into the zone of absorption. Applied to the soil in soluble form, arsenic may be leached into the root zone where it exerts its toxic action. However, on a concentration basis, it is much less effective applied through the soil than it is applied in culture solution. This indicates that its toxicity to plant roots is greatly reduced through con- tact with the soil. Arsenic, like phos- phate, is fixed by the soil, and much of the fixed arsenic is rendered unavailable to plants; a small proportion remains immobile but available. It is this frac- tion that is important in soil sterilization. Work with 80 California soils proved that arsenic toxicity is determined largely by textural grade as this char- acter relates to colloidal content. Kao- linitic clays are particularly effective in fixing arsenic in a form unavailable to plants. Recommendations for use of ar- senic trioxide in soil sterilization in Cali- fornia call for application during the winter months when rainfall or snows carry the chemical into the topsoil. The object of this recommendation is two- fold: (1) to get the arsenic into the moist soil so it will start dissolving; (2) to make the application when there is little or no vegetation on the soil, thus minimizing the poison hazard. All ap- plied arsenic is subject to leaching by rains or melting snow; however, arsenic resists leaching and is washed from the soil only by large quantities of water not normally falling as precipitation. Ar- senic applied as dry crystalline trioxide is less readily leached from the soil than arsenic applied in solution. An important objection to the use of arsenic in weed control has been its great poison hazard. This is particularly true of soluble arsenic, which is not only poisonous but also attractive to livestock. Arsenic trioxide apparently has no such attraction, and its use has proved safe even in areas in which livestock are grazing. At the California Forest and Range Experiment Station arsenic tri- oxide was spread on firebreaks in open range country where livestock and wild- life were present. The animals were curi- ous ; they came up and sniffed the arsenic but did not taste it. The present program is aimed at de- veloping the use of arsenic trioxide for soil sterilization along railroads, irriga- tion ditchbanks, fence lines, rights-of- way, airfields, firebreaks, and other places where all weed growth is unde- sirable. [4] THE STUDIES . . . Three methods were used to study the herbicidal properties of arsenic trioxide: (1) greenhouse tests on arsenic toxicity in soils, using the concentration series; (2) field plots to find the effects of arsenic application on various weeds growing in several soils and exposed to differing environmental condi- tions; and (3) applications on a field scale to an airfield, to a series of structures on a large irrigation system, and to fire- breaks in forest and range areas. Greenhouse tests The experiment. The toxicity stud- ies follow the pattern that has been used in the studies on sodium arsenite (Crafts, 19356; Crafts and Cleary, 1936; Crafts and Rosenfels, 1939), sodium chlorate (Crafts, 1935a, b, c; Crafts, 1939a, b; Rosenfels and Crafts, 1939), borax (Crafts and Raynor, 1936; Crafts, 19396), and other herbicidal chemicals. Small pot cultures were set up in series, No. 2 cans (unperf orated) serving as containers. The concentration series used in these tests runs as follows: 0, 40, 80, 140, 220, 340, 490, 680, 920, 1,220, 1,590, 2,040, 2,580, 3,220, 3,920, and 4,840 ppm As 2 3 in air-dry soil. Weighed amounts of arsenic trioxide were mixed with 500-gram lots of soils by rotation in closed jars on a soil mixer. The dry mixtures were moistened to field capacity by adding water to weight. The cultures were planted using 13 Kanota oats per can; one week after germina- tion they were thinned to 10 plants per can. They were harvested 30 days after planting, and the green weight and aver- age height for each culture were re- corded. All series were run in triplicate. Figures 1 and 2 show some of the cul- tures at harvest time. In addition to the straight toxicity series, test series were run in which pe- riods of 2, 3, 5, 7, 9, and 16 days were 1 \% li ii if/ \H \ [■//* Jl" ; -' / \\ wi \\m .\\M// ■ ti II +m&Ai"Av?z. 1 milm m 1 ■ < ■ : - ? M w s H - m 1 - I ■ *' f " 1 H ■ M m m 1 mi m k*r 1 <*t . '""■' mitim- | h WP^' isp JlwJ Fig. 1. Toxicity of white arsenic to crop #1 in two California soils; left, Sierra fine sandy loam, and right, Aiken clay loam. Toxicity is highest in sandy soils. In both series the cultures on the left are controls, the center cultures have 490 ppm in the air-dry soil, the right-hand cultures 1,220 ppm. [5] Fig. 2. Left, toxicity of white arsenic to crop #1 in Hanford loam, and right, Yolo clay loam. Left, (in each photo) controls; center, 490 ppm; right, 1,220 ppm. allowed to pass between moistening the cultures and planting the oats. This was to allow time for solution of the arsenic trioxide to take place. Yolo clay loam was used in this experiment. Two toxicity series were established in which a wetting agent (Tenlo 400) was included in the moistening water. These were run in Hanford fine sandy loam and Yolo clay loam and were cropped for three 30-day periods with intervening 30-day drying periods. In field-plot tests the dustiness of the dry arsenic proved to be a nuisance and a hazard to the operator making the ap- plications. To avoid this and to mini- mize blowing of the arsenic where ap- plications were being made to railroad ballast, about a ton of arsenic trioxide was pelleted for experimental applica- tion. For similar purposes a small quan- tity of a granular form of As 2 3 was supplied by the American Smelting and Refining Company. Greenhouse trials were run using plain arsenic trioxide dust, pelleted arsenic trioxide, and the granular form. As an alternative method to reduce dustiness, liquid wetting agents were added to the dry dust. This proved very effective, and samples of several such mixtures were supplied by the California Spray Chemical Company for testing under greenhouse conditions. Data on these tests, which were cropped for seven successive periods in Yolo fine sandy loam, are given. Table 1 characterizes the soils that were used in the greenhouse tests. The results. Data on six runs of the toxicity series are presented in figures 3 to 12. Figure 13 presents the results of similar tests with sodium arsenite in a group of 80 California soils (Crafts and Rosenfels, 1939). A comparison of the results shows that in the first run toxicity is much lower with arsenic tri- oxide than with sodium arsenite. This is to be expected because arsenic tri- oxide is low in solubility and requires time to dissolve in sufficient quantity to provide a lethal concentration in the soil. Though somewhat less apparent than when soluble arsenic is used, the relation between toxicity and textural grade is evident in these results. By the time the second run was har- vested, toxicity had increased appreci- ably in each soil. This is in line with results in the field where it has been ob- [6] Table 1. Characteristics of California Soils Used in Toxicity Tests Soil type Origin Mode of formation Stage of development Color Reaction Aiken clay loam Basic igneous Primary Semimature Red Slightly acid Egbert loam Mixed Secondary Immature Dark Slightly organic alluvial gray acid Hanford fine sandy loam. . . Acid Secondary Recent Light Neutral igneous alluvial brown Hanf ord loam Acid Secondary Recent Light Neutral igneous alluvial brown Rocklin fine sandy loam . . . Acid Secondary Mature Brown- Slightly igneous alluvial red acid Sierra fine sandy loam Acid Primary Semimature Brown- Mod. igneous alluvial red acid Stockton adobe clay Basic Secondary Mature Gray- Basic igneous alluvial black Willows adobe clay Sedimentary Secondary alluvial Semimature Dark brown Neutral Yolo fine sandy loam Sedimentary Secondary alluvial Recent Brown Neutral Yolo clay loam Sedimentary Secondary alluvial Recent Brown Neutral served that sterility does not reach a practical level in medium to heavy soils until the second season. In light sandy soils sterility may be very evident the first season, particularly on soils low in fertility. In most of the soils maximum toxicity is not attained until the third run. Sierra fine sandy loam is the only exception; in this soil the second run exhibited maximum toxicity. This is explained by the coarse granitic nature of the sand in this soil. Runs number four and five were not significantly different from run number three in these soils, but in several soils run number six showed loss of toxicity. This is understandable in the case of Stockton adobe clay and Aiken clay loam on the basis of fixation by clay col- loids, but in Sierra and Rocklin fine sandy loam this reasoning does not ap- ply. Kaolinitic clays in these soils may account for the lowered availability of the arsenic. Figs. 3 to 12 show the relation of crop yield to the concentration of white arsenic in the various soils tested. The graphs for these figures appear on the following pages. [7] Z.? 2 2 5 ! »^_ "--> AIKEN CLAY LOAM Run No. 1 Run No. 2 Run No. 3 Run No. 4 — — Run No. 5 Run No. 6 — 1,000 2,000 3,000 p. p.m. As Qj in air-dry soil Fig. 3 4,000 5,000 HY\/\ EGBERT LOAM Run No. 1 Run No. 2 Run No. 3 Run No. 4 Run No. 5 Run No. 6 - _ — *V^ ^ " = "-■■— *u-^. — — ^:- ^r-_-iv.-z — 1,000 2,000 3,000 p. p.m. As ? 3 in air-dry soil Fig. 4 4,000 5,000 2,000 3,000 p. p.m. As 2 Qj in air-dry soil Fig. 5 [8] 1,000 2,000 3,000 p. p.m. As ? 3 in air-dry soil Fig. 6 ROCKLIN FINE SANDY LOAM Run No. 1 Run No. 2 Run No. 3 Run No. 4 Run No. 5 Run No. 6 3,000 5,000 Fig. 7 1,000 2,000 3,000 p. p.m. As 2 3 in air-dry soil 4,000 5,000 Fig. 8 E 6 5 -, * v \ ****^i- STOCKTON ADOB E CLAY Run No. 1 Run No. 2 Run No. 3 Run No. 4 Run No. 5 Run No. 6 ""-"" — " — "S'^ir-s-iH ^/iP" .— _--— ^-_— -i.— -^ — ^ - ------ u — ---_ 1,000 2,000 3,000 p. p.m. As O^ in air-dry soil 4,000 5,000 Fig. 9 [9] i O) V ~V\ ^v^ WILLOWS ADOBE CLAY O) 1 4 U- to 2 ^ Run No. 1 Run No. 2 Run No. 3 Run No. 4 Run No. 5 Run No. 6 < CL f) «««!«« ===-« 1,000 2,000 3,000 p. p.m. As 2 3 in air-dry soil Fig. 10 5,000 2,000 3,000 p. p.m. As 3 in air-dry soil Fig. 11 5,000 2,000 3,000 p. p.m. As 3 in air-dry soil 4,000 Fig. 12 [10 1 100 — o Sands - —■— ° Gravelly and sandy loams -o Loams o— — — o Silt and clay loams o o Clays V X ^N; \ is. *^8Trs, 200 E3 400 600 p.p.m. As 2 3 in air-dry soil 800 1,000 Fig. 13. The relation between textural grade and arsenic toxicity from the summaried results of toxicity tests in 80 California soils. (From Crafts and Rosenfels, HILGARDIA, Vol. 12, Calif. Agr. Exp. Sta.) Figure 14 presents the data on the ex- periment in which periods of from 2 to 16 days were allowed between moisten- ing the cultures and planting them. It is apparent from the values that periods up to and including 9 days resulted in no significant increase in toxicity. Fig- ure 14 shows three curves: (1) the first run in Yolo clay loam converted to a control value of 11.2, (2) a curve made by combining the values for the 2-, 3-, 5-, 7-, and 9-day delay periods, and (3) a curve for the 16-day delay period. Since the culture series ran to only 1,220 p.p.m., the curves do not go to the base line. However, the trends are clearly shown, and there is no doubt that the 16-day period has been long enough to allow appreciable solution of the arsenic trioxide in the soil. Hanford fine sandy loam and Yolo clay loam were used for the experiment with wetting agent (Tenlo 400) . Figures 15 and 16 show the results. In the Han- ford soil the differences were hardly significant; in Yolo clay loam the tox- [in 1,250 Fig. 14. Toxicity of white arsenic in Yolo clay loam as related to solution time for the applied arsenic. 5^==- HANFORD FINE SANC r )Y LOAM Run No. 1 Run No. 1 with i Run No. 2 - — Run No. 2 with \ vetting agent ~~\ \ \ — . ~~"^\^ ., ~~ ■"" — ~ --^ 1,000 4,000 Fig, 2,000 3,000 p. p.m. As 2 3 in air-dry soil 15. Toxicity of white arsenic in Hanford fine sandy loam as affected by addition of a wetting agent (Tenlo 400). 5,000 1,000 4,000 2,000 3,000 p. p.m. ASjOj in air-dry soil Fig. 16. Toxicity of white arsenic in Yolo clay loam as affected by addition of a wetting agent (Tenlo 400). 5,000 4 ^\ r -0~~^*v YOLO FINE SANDY LOAM Run No. 1 2 n — «^ """~~ ~~ — — — , -—-- — ___ V s *" "~ ~~ - \ X ^ Run No. 2 \ "U^ Dust Granular Pellets 1,000 2,000 3,000 p. p.m. As O in air-dry soil 5,000 ^N \ Run No. 4 \N N -X s * ^ X s "*>• \ N ** \ \ ^ ^» ""^ — ~~ — — — — . — =_ — 6 \ Run No. 7 4 2 ■ V ^ X x \ \ 1,000 2,000 3,000 p. p.m. As 2 3 in air-dry soil 4,000 5,000 Fig. 17. Toxicity of white arsenic in Yolo fine sandy loam as related to the form applied. Four runs with dust, pellets, and a granular glassy form. icity of the arsenic was materially in- creased, run number one with wetting agent approaching run number two without, and run number two with wet- ting agent being more effective than run number three without. Whether the ef- fect is due entirely to greater solution of the arsenic or to a more complex response involving the wetting agent as an inhibitor of fixation cannot be de- termined from these simple tests. They do indicate, however, that inclusion of a wetting agent to reduce dustiness may also increase toxicity during the first year after application. When the question of pelleting ar- senic trioxide was raised, representa- tives of the American Smelting and Refining Company suggested the pos- sible alternative of using the granular [13] arsenic that is produced along with the dust in the smelting process. Trials were run with this granular form and with some pellets that had been pro- duced experimentally, the dust serving as a control. Figure 17 presents the re- sults of this test in Yolo fine sandy loam. Toxicity of the granular form and the pellets was very low in the first run; in the second run it was consider- ably higher for each. In the fourth and seventh runs the pellets were as toxic as the dust form, and by the seventh run the granular form was approaching the other two. Granular arsenic is evidently very low in solubility; even by the seventh run it had released only as much soluble ar- senic as was available from the dust dur- ing the first run. Arsenic in the pellets was somewhat slower to dissolve than the dust. Furthermore, the pellets were dispersed through the soil mass in defi- nite restricted locations during the first run, and the roots of the oat plants could grow through the unaffected soil be- tween them. By the second run, arsenic from the pellets was somewhat more thoroughly mixed with the soil but still not completely. Possibly also the binder used in the pellets restricted solution. That this is not serious in field applica- tions is indicated by the high toxicity of the pellets in the second year of treat- ment. In general, the pellet applications were as effective as, or in some cases more effective than, dust applications on railroad ballast. In some situations this might have reflected the greater stability of the pelleted form with respect to wind action from passing trains. While the above trials were being con- ducted, one company, interested in the sale of arsenic trioxide as a soil sterilant, submitted four samples that had been formulated with different wetting agents to produce dustless products for com- mercial application. Run No. 2 4 C N -\ 1,000 2,000 3,000 p.p.m. AsjO, in air-dry soil 5,000 Fig. 18. Toxicity of white arsenic in Yolo fine sandy loam as related to formulation. A compari- son of dust, pellets, and four formulations involving mixtures of commercial surfactants with arsenic dust. [14] ■\i :r fi Fig. 19. Sterilization plots seven years after treatment with arsenic trioxide. The heavy lines mark 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. Figure 18 gives the data for runs num- ber one and two with these materials. Lot number 2,126 was little better than the dust during run number one; lots 2,124, 2,125, and 2,127 were considerably better. By the second run all formula- tions and dust and pellets gave compa- rable results. These tests indicate that considerable time can be gained in ob- taining sterilization with arsenic trioxide by incorporating a wetting agent. Evi- dently, however, these agents differ somewhat in their solubilizing power. During the time the above tests were being made, the new organic soil steri- lant, p-chlorophenyl 1, 1, dimethyl urea (CMU), was presented for trial. Com- binations of arsenic trioxide and CMU were tried in hope of finding a mixture that would be effective in the year of application. These tests proved that there is little or no antagonism between ar- senic trioxide and CMU, that the effects were additive, and that such a combina- tion is very toxic, provides rapid sterili- zation, is effective the first year, and lasts as long as the arsenic dosage de- termines. Plot tests The experiment. Plot tests with ar- senic trioxide were established at Davis in October, 1931 (Crafts, Bruce and Raynor, 1941, table 1) . Since then many hundreds of trials have been made. The object was to establish the dosage re- quirements under widely different cli- matic conditions and to learn more about handling the chemical in the field. Dosage requirements under various climatic conditions were studied by es- tablishing plots on railroad main lines and sidetracks at various locations from Marysville on the north to Lodi on the south. To include altitude as a factor, series of plots were established at Rose- [15] Fig. 20. Left, plots treated with white arsenic. Released from competition with annual weeds, Elymus triticoides, a deep-rooted perennial grass, has flourished and now almost completely occupies the area. A readily leached soil sterilant is required in addition to white arsenic in this situation. Right, ballast and tow path sterilized with white arsenic. A patch of E. triticoides is shown encroaching on the tow path on the left of the track. ville, New England Mills, Colfax, Towle Station, Yuba Pass, Soda Springs, Truckee, California, and Verdi, Nevada. On this latter group of plots arsenic dust and arsenic pellets were compared. A pushcart applicator was used to spread the chemical. Three years' results have been observed on these plots. The results. Figure 19 shows one of the early plots established for a previous study (Crafts, Bruce, and Raynor, 1941). The object of the plot studies in the present project was to improve ap- plication methods and to gain practical experience in handling arsenic trioxide in the field. The first series was treated in February, 1950, and consisted of 16 plots, each two square rods in area, lo- cated in pairs. One had a west exposure and extended 32 feet from the tracks down into the borrow pit of the Southern Pacific Railroad where the latter tra- verses the Armstrong Tract of the Uni- versity of California south of Davis. The second plot of each pair faced east on the same right-of-way. Treatments in- cluded applications of 2, 4, 6, 8, 10, 14, 16, and 20 pounds per square rod. The chemical was applied when the vegetation was from 6 to 12 inches high. A second series of plots was treated with applications of 2, 4, 6, 8, 14, and 20 pounds per square rod in March, 1950, when the vegetation was 10 to 14 inches high. There was little visible effect on any of these plots during the season of treatment. Examination of these plots in the early summer of 1951 and again in 1952 proved that the arsenic application re- quired to control annuals on this Yolo loam soil on sloping terrain was 10 pounds per square rod or more. And where the deep-rooted grass Elymus tri- ticoides was present, it thrived and com- pletely covered the treated plots. Figure 20 shows such a plot. This stresses the fact noted in previous studies (Crafts, Bruce, and Raynor, 1942; Raynor, 1937) that arsenic trioxide is not effec- tive against perennial weeds and should not be used alone in situations where such weeds occur. Also noted was the fact that sloping soil is more difficult to sterilize than level soil. This has a prac- tical bearing in many situations where soil sterilization is desired. In November, 1950, a series of arsenic plots was established on the Southern Pacific Railroad along the main line [16] south of Davis. A second series was treated in December, 1950. In each, ar- senic trioxide, both in dust and pellet form, was applied with a three-foot pushcart applicator, designed and built by Norman Akesson of the Agricultural Engineering Department at Davis (see page 2). Table 2 presents the readings on these plots, taken in May of 1951 and 1952. It is evident that the 1952 readings represent a more complete sterilization than those of 1951. This is typical of arsenic trioxide and confirms the greenhouse tests where maximum toxicity was not attained until the third cropping on some soils. A second conclusion that may be drawn from the data in table 2 is that the pelleted arsenic was somewhat more effective than the dust. These plots were on the main line over which many trains passed daily. It was noted that the pas- sage of trains during and immediately after the applications caused the arsenic Table 2. Results of Soil Sterilization with Arsenic Trioxide on Railroad Plots (Ballast and Tow Path) South of Davis.* Application by Mechanical Applicator (Three-foot Pushcart). Notes taken in May, 1951 and 1952 Arsenic form and date of application Application Percentage of control 1951 1952 Dust pounds per acre 500 750 1,000 1,280 1,500 1,750 2,000 30 45 60 65 60 65 75 60 75 85 85 90 90 98 (Oct. 28, 1950) Pellets 500 750 1,000 1,280 1,500 1,750 2,000 35 50 60 60 70 65 75 70 80 90 95 98 95 98 (Oct. 28, 1950) Dust 750 1,000 1,280 1,500 1,750 2,000 25 45 65 70 70 90 25 75 85 80 90 90 (Dec. 10, 1950) Pellets 750 1,000 1,280 1,500 1,750 2,000 50 70 80 80 90 90 80 90 90 90 98 99 (Dec. 10, 1950) * Plots 9' x 100'. [17] dust to rise and blow in swirling cur- rents, some actually being blown off the plots by each passing train. This did not take place where pellets were used. In fact, close inspection proved that these tended to settle in crevices between the rock ballast and so to concentrate in the regions where seeds tended to germinate. This is a distinct advantage. In November, 1950, a series of plots was established on the Western Pacific Railroad between Sacramento and Lodi. These are reported in table 3. Most of the plots were located on mounds around signal towers. These mounds are graveled or paved with crushed rock and are difficult to keep clean by hoeing. By 1952 the arsenic had done an excellent job of sterilization against annual weeds, but in many cases perennials persisted and where they were not killed by spray- ing they tended to invade the areas. Of the annual weeds that were able to gain entrance to the areas, yellow star thistle (Centaur ea solstitiales) and knotweed (Polygonum aviculare) were the two most serious. During December, 1950, and January, 1951, additional plots were established near Woodland in Yolo County and at Bahia Station in Solano County, all alongside tracks and spur tracks of the Southern Pacific Railroad Company. Table 4 presents the readings on these plots, and figures 21, 22, and 23 show the results. A final series of plots on the Southern Pacific main line was established in De- cember, 1950, and January, 1951. These plots were located from Roseville through the Sierra to Verdi, Nevada. They covered wide ranges in natural pre- cipitation and altitude and included areas where snow is prevalent through- out the winter. These plots are reported in table 5. Figure 24 shows the treated Table 3. Results of Soil Sterilization with Arsenic Trioxide Dust, Applied November 3, 1950, on Railroad Plots between Sacramento and Lodi. Application by Hand Spreading. Notes Taken in May, 1951 and 1952 Size of plot Location Applica- tion Percentage cf control Notes 1951 1952 feet 150 X 12 25x25 W.P. RR. at Alberta pounds per acre 2,000 2,000 80 75 98 98 Winter annuals all dead ; yellow star thistle encroaching 25x25 100x6 25x25 W.P. RR. at Kingdom Sta. 2,000 2,000 2,000 65 80 75 75 95 90 Yellow star thistle encroaching ; also perennial weeds. 300x6 Terminous road 2,000 75 95 Yellow star and mayweed 25x25 25x20 Thornton, Cross- ing 113.3 2,000 1,000 75 50 98 95 Knotweed and yellow star encroaching 25x25 South of Franklin, Crossing 118.6 2,000 75 95 Yellow star encroaching 25x25 Franklin 1,000 90 98 Poppies and milkweed en- croaching [18] 0) .d +» • d o c in CO •♦» CO eo ! 4> oi 01 5 .0 CO •d •a §1 5 * £ ■S o £ S H N CO *d CO 6 23 ■S o CO *-. ^j (WO odd fc fc * -^ •* ■* o o CQ CB CO |m| k| J-i >-i t-i +j -u 9 S 9 1 1 CO bo f c0 CO o8 CO h CO bo 1 cO ■d H ■d CO CO CO > 4 1 _ 0Q ft a an h w w w s <; H o S.S o c ^ c § o 3 UO o o o fr- UO UO lO © © © © uo 05 05 o OS 05 (A J* o o s fl) fli C9 00 t* 05 05 o 05 05 -2 ■- *o a. (A _ o ho ' ■* J o a o Z © to o o o o U0 Tj< -^ CO UO O lO o o o o UO O uo M 3 t» CO t- t» CO 00 00 oo 00 00 E t a o Si 1 | o 5 8 8 8 8 ft N N N N II III II o s || •™ 4- ^. •O rl H H rl X o p. s < § H- O Pi •d ■2 .2 ° 1 d cO 53 1 d) C O o n with Ars pplicator (' J o o ■3 & » & *© ? © s o 2? § SP © d a. 3 ►> CQ w eo J> CO CO "c5> to ■8 § Pi 1 "S o o o 1 o in cO 1 n .2< 4- — 8 .- .s © CN CM o IE "£ w 8 .3 ta ^ H ^Ji 00 00 | X X X X CD CO (O CD CD X X X X X CO 05 X X X UO CO X X J>X s III! o o o o o O uo o o o 83 o o S8 «o s 00 CN CO 00 00 i-H M CO to >. T-T •A M- J* o c 02.2 s s o UO o UO uo 3 8 «a 3 1 rH 03 8 05 T-l t> CN 05 t> CN 05 o 3 d d d d FJ 0) 4} CO CO Q P Q Q 4 1 8 1 1- ii CO 3 1 M o +» +» ■*» ■+» +» *» 4J CO +» CO +» +* ■<-> -t^» Js*** co co ta to CO CO CO ^ CO « d to CO CO <5 333S S3 d 53 » d 2 d d PQQQ QQPBhP £ Q 1 Q 1 O Q strips at Truckee, Calif. The data bear out the observations that sterilization is better after the third winter season, that the pellets in general are superior to dust, and that annual weeds succumb to the treatment but perennials usually sur- vive. The notably poor results at Verdi in May, 1951, probably reflect the low total precipitation in this area. These contrast with the generally better results on the wetter west slope. In conclusion it should be emphasized that the data presented in tables 2 to 5 relate only to annual weeds. If readings on perennial weeds or total weed growth had been reported, they would be dif- ferent. This proves that one chemical cannot be expected to stay in the top- soil and control seedlings and annual weeds, and at the same time leach into the lower soil horizons and act on the roots of perennials. Wherever annuals and perennials occur, soil sterilization involves two separate processes and, under most conditions, requires two chemicals. One of these, because of fixa- tion or low solubility, should remain active in the surface soil, while the other moves into lower levels to handle the deeper-rooted perennials. Arsenic is pri- marily a surface sterilant. Only when applied in water-soluble form (sodium arsenite) to very sandy soils does it pen- etrate deeply enough to kill perennial weeds. Arsenic trioxide, because of its relatively low solubility and also because of fixation on soil colloids of that which does dissolve, remains in the topsoil. Its virtue is its long-lasting toxicity against seedlings in the topsoil. It should Fig. 21. Bridge on Southern Pacific track cleared of vegetation by soil sterilization. Fig. 22. Ballast of a feeder line sterilized with white arsenic. Prior to treatment this line was heavily infested with yellow star thistle. Fig. 23. Spur track in the yards of the Spreckels Sugar Co. at Woodland sterilized with white arsenic. The photograph shows the end of a treated strip and emphasizes the weedy nature of the area. be mentioned here that seedlings of per- ennials are like those of annuals — that is, they are susceptible to control by sur- face-acting sterilants. It is the old, estab- lished perennials that require a deep- acting soil sterilant. Field-scale applications Irrigation structures. The first field-scale work on this project was done by the U. S. Bureau of Reclamation at the suggestion and under the supervi- sion of the University of California. In some 60 treatments arsenic dust was ap- plied to bridgeheads, cattle guards, and other structures along the Delta-Mendota Canal. These applications were made during January and February, 1950. In November, 1950, a series of tests was made in which arsenic trioxide was ap- plied as narrow strips along the lining of the same canal. These were put on by the University working with the Bu- reau of Reclamation. During this same winter period further treatments on structures along the canal were made by the Bureau of Reclamation. In the early summer of 1951 inspection proved that the treatments of January and Febru- ary, 1950, were giving almost complete control of annual weeds. Since the canal traverses range land and relatively new agricultural land, there are not many perennials present; the treatments have proved very satisfactory. Even the treat- ments of November, 1950, gave excel- lent control by May, 1951, possibly because of heavy rains during the winter and also because the soils treated tended Fig. 24. Soil sterilization along the right-of- way of the Southern Pacific mainline through Truckee, Calif. Arsenic trioxide at 1,000 pounds per acre. Fig. 25. Soil sterilization around a cattle guard on the Delta-Mendota Canal. White arsenic was used. Courtesy Bureau of Reclama- tion, Region 2. Fig. 26. White arsenic sterilization along the lining of the Delta-Mendota Canal. Courtesy Bureau of Reclamation, Region 2. *jSP **'vX -:•?■>* q.- f X s :-v • -:*\. ■u « c 10 o «> •S il CO CO ;£ 1 d u C ca o CO _ u ■- DO o & CO > o CO »d d CO a o PI o Si ed CO een Roseville, Taken in June, fc bo bo •Bag Sis s <=» s ^ >h O c3 ft) 3* I 0)10 lO 00 lO O 00 OS IO IO io oo IO OS 8 iH OS IO o u s OS OS t- OS OS cs OS OS OS OS !> OS OS OS OS OS OS «A • *« It o bo a 8 5 u d -D -= o io © o o o O IO IO O O lO o IO IO o o o 1 S 2^ a ooosio IO IO os IO t- CO t- lO t> oo t- 00 OS IO IO rz *■ 5 © 0) £ o J I n O O O o o o o o o o o o o o o o o o c o I £ O O O 00 00 o N NO 00 00 o 00 00 o o o o o 00 o CN O Pi IO IO IO CN CN O CN CN o o^ o o_ ©^ k "3. •8 rT H H rH iH CN rH~ iH CN rH CN O X p. ■u t, <1 O n ■- ft o ■d 6 E* w — " 8: g ■a 3 £ r« * I '3) CO •d cd d O I 4» CO CO cfl Ph efl .o S3 CO be CO iH s >> s-3 CO CO o > co* g * > s d CO W !=! * IS «0 ^ >> o 02 CO I— t o Pi 02 ca o CO CO 1 •d "9 6 S .t c rt fc o H 5H 02 Eh rH ^ c if II .3 ^ CD (O U) IO OS CO CM IO CO CO CN CO CO CO CO CO CO OS co os s* ., g XXX XXX XXX X X X X X X X X X O *" o o o CO CO IO M H lO CO o CN lO os os co co OS CO IO — x Ah O IO o N IO N CO 0) CN CO CN O tH T-i 00 lO co tj< aM H N rl rH rH "* ^ CO CN CN iH tH rH u c a. 2 4- _ 8 •— u O — u> a tH M- OL o < IO OS o IO g § s o IO s g e9 rH OS OS os OS os os OS «A • & tH rH iH rH tH rH rH ±Z o CN IO IO IO IO co CO co" 3 "D . . . . in o *s > 1 o CO o CO eo CO CO CO CO CO d CO d CO p P P p p p P 4) :z 11 CO CO CO CO CO — ■ T5 4» 4» 4* ■** 4» X *- <> 2 o 4* +» 4» 43 CO 4* 4* CO 4» ■u CO ■*» CO -M CO •fa ■U 49 ■$T CO CO CO a s s CO ^ CO d » 3 CO ^ to co sj 3 Is d CO P co CO d CO CO d d h- > QQQ QhQ Q PL, Q P Ph p ft* P Ph P PP to be gravelly or sandy. Figures 25 and 26 illustrate sterilization on cattle guards and along the lining on the Delta-Men- dota Canal. Airfield. A second large-scale treat- ment involved application of about 6 tons of arsenic trioxide along the edges of the flight strips of the University Air- port at Davis. This material was spread by means of a converted fertilizer drill that covered an 8-foot strip. Put on in late November, 1950, this arsenic has greatly reduced weed infestation which threatened to break up and ruin the thin oil-gravel pavement that is used to pro- tect the flight strips from invasion by weeds. Experience at this airport has revealed one serious problem in the use of arsenic trioxide, that is, the persistence of cer- tain arsenic-resistant weeds. Here knot- weed (Polygonum aviculare) has proved the most difficult to handle. Though literally millions of seedlings of this plant have been killed each year on this sterilized area, each season a few seed- lings survive, and, being free of com- petition, they thrive and grow to a surprising size. One of these plants growing through a small crevice in the pavement may spread to a diameter of 6 to 9 feet, making a dense bush a foot or more thick in the center. The root may be fully one-half inch in diameter, and when the plant dies and disinte- grates in the winter this hole provides entry for more plants the next year. Yel- low star thistle (Centaur ea solstitiales) is another weed that tolerates a surpris- ing amount of arsenic. Where these weeds survive the arsenic treatment, spraying with Endothal has kept them in control. Firebreaks. A third field-scale use of arsenic, the largest so far made, has been the treatment of many miles of fire lines in the mountain regions of California. These include southern Cali- fornia firebreak systems and cleared Fig. 27. Left, applicator for treating firebreaks with white arsenic. Two men can treat several miles of backfire line in one day; the soil remains sterile for 5 or more years. Right, treated strip •showing the thin, even coat of white arsenic. During rainfall the white arsenic is washed into the soil. Even when first applied, it holds no attraction for livestock. Courtesy H. D. Bruce, Calif. Forest and Range Expt. Sta., U. S. Forest Service. [23] Fig. 28. Left, fire line at edge of northern California railroad right-of-way still bare nine years after treatment with 5 pounds of white arsenic per square rod. Right, it is necessary to spread arsenic by hand where the terrain is too rough for mechanical equipment. This picture was taken 9 years after treatment with between 5 and 6 pounds per square rod of soil. Courtesy U. S. Forest Service. strips along roads, railroads, and power- lines in both north and south, created to confine fires starting on them to the right-of-way. Both hand and machine applications have been made, although machine- spreading has accounted for most of the treated mileage. Hand-spreading has generally been confined to spot treat- ments around individual poles on power lines and to comparatively short, al- though no less important, stretches of fire line too rough for mechanical spreader operation. Equipment similar to that shown in figure 27 (left) has been used on most of the treated miles. Wherever erosion is not serious and the soil has remained undisturbed after treatment, the effects of both hand- and machine-spread arsenic have been long lasting. Figure 28 shows two sections of a hand-treated fire line along a railroad right-of-way in northern California nine years after treatment. Five pounds of arsenic per square rod was used. Four pounds of arsenic per square rod spread by machine on a southern California fire- break has been almost equally effective against plant invasion (figure 29). Note the burned vegetation on both sides of this firebreak. The fire occurred about four years after treatment. In spite of the barren strip firefighters were unable to stop the fast-running fire at this point. The vegetation on both sides of — and in- side — the treated line has seeded in since the burn. This line was disturbed by heavy traffic during the fire. Spreading arsenic by hand often re- sults in a heavier application than neces- sary because hand-spreading cannot give complete ground coverage with as thin a coating as can be obtained with the mechanical spreader. Nearly complete coverage is necessary because of the ex- tremely limited lateral movement of the arsenic. Some of the reasons uniform coverage with minimum quantities of chemical cannot be achieved with hand-spreading are apparent in figures 28 (right) and 31. In a boulder patch such as that in figure 28 (right) , one has to work around each rock, throwing arsenic in under each overhanging edge. The cut slope and a 3-foot uncleared strip at its top (figure 30) were sterilized to eliminate cheatgrass as a fire-prevention measure on this busy mountain highway. Dead grass remained on the uncleared strip until the second year after treatment. After that the ground was bare. [24] Fig. 29. Heavy traffic by firefighters scuffed up the soil on the treated portion of this south- ern California firebreak. Even with this dis- turbance, invasion by perennials 7 years after treatment with 4 pounds of arsenic per square rod has been slight. Courtesy U. S. Forest Service. Whether chemical or mechanical methods are used, permanent removal of vegetation is not always desirable. Ero- sion is probably the most costly result of misapplication by either method. Veg- etation on the firebreak shown in figure 31 was cleared two years before the pic- ture was taken. Gully erosion has resulted in excessive soil movement over much of the width of this firebreak. This strip, which traverses an area of decomposed granite, is too steep to remain stable when bare without special provision for drainage. Arsenic applied to an area such as this will usually disappear en- tirely by the fourth year. A few water barriers on this slope would have pre- vented this damage and made arsenic treatment practical. Experience on both firebreak and right-of-way sterilization has shown the need for posting the treated ways against damage by bulldozers and motor-patrol graders. These bare strips often provide a convenient path for such equipment, and operators frequently let the blade down just to smooth the way as they go. Since arsenic trioxide becomes fixed in only the very top soil layer, blading invari- Fig. 30. Treating the cut slope and the 3-foot wide strip at the top along this northern Cali- fornia highway. These areas are being sterilized as a fire-prevention measure. At least two years is usually required to get a bare strip without pre-clearing. Courtesy U. S. Forest Service. [25] Fig. 31. This southern California firebreak in decomposed granite was too steep for success- ful sterilization without adequate provision for drainage. Erosion such as this, when once started, is difficult to stop. It is less severe than that which often follows disking or hand- hoeing, for these mechanical operations loosen the topsoil. Courtesy U. S. Forest Service. taken to apply it on relatively calm days, ably renders the treatment void. Posting the treated strips with appropriate warn- ings seems to be the only solution, since turnover in both supervisory and opera- tor personnel makes verbal instruction of only temporary value. When the fine granular form of ar- senic trioxide is used, care should be as wind will carry appreciable quantities of the chemical off the treated area. Until the chemical is wetted, it is also suscep- tible to wind erosion after it is on the ground. Most extensive treatments have, therefore, been carried out in the late fall and spring months between rains. Where work programs have made treat- ment desirable during the dry season, a light sprinkling of treated strips by tank truck has been equally successful. It is even much better than a very heavy rain immediately following application, as this may wash some of the arsenic away. Many public agencies and utility cor- porations are averse to broad-scale use of arsenic. Their attitude stems pri- marily from its bad reputation estab- lished several decades ago by applica- tions of arsenic in the form of sodium arsenite. Both arsenic trioxide and sodium ar- senite are poisonous to all animals if taken internally. While sodium arsenite, however, is salty in taste and attractive to grazing animals, arsenic trioxide is unattractive to them and is therefore not a hazard to them once it is on the soil. No harmful effects from soil steriliza- tion treatments with dry arsenic trioxide have been reported in more than ten years of use. Arsenic trioxide is only slightly sol- uble in water and is therefore usually spread on the ground dry. Buyers of arsenic trioxide for weed con- trol are cautioned against it in solution form. Hydroxide is used to dissolve arsenic trioxide, and this converts it to the dan- gerous arsenite. Arsenic trioxide should be used dry as dust, dustless powder, or pellets. In order that the information in our publications may be more intelligible it is sometimes necessary to use trade names of products or equipment rather than complicated descriptive or chemical iden- tifications. In so doing it is unavoidable in some cases that similar products which are on the market under other trade names may not be cited. No endorsement of named products is intended nor is criticism implied of similar products which are not mentioned. [26] THE ANSWERS . . . Arsenic trioxide is the most economical chemical to use where sterilization of 5 to 10 years' duration is desired. The grade known as gray arsenic, assaying 90 to 95 per cent As O , is satisfactory for soil sterilization purposes. Here are the con- clusions in detail. Toxicity In 10 California soils arsenic trioxide was less toxic than soluble sodium ar- senite during the first cropping of a soil. By the third cropping its toxicity was at a maximum and equal to that of sodium arsenite. By the sixth it was diminishing. As with arsenite, arsenic trioxide toxicity was highest in light sandy soils, lowest in heavy clays. When arsenic trioxide stood in contact with moist soils for 2 to 9 days, no in- crease in toxicity resulted; after 16 days an appreciable increase could be de- tected. In the powdered form arsenic trioxide has little tendency to wet in water. Addi- tion of a wetting agent brings about ready wetting, and in one soil this in- creased wetting was attended by an increase in toxicity during the first cropping. Added at 0.1 per cent of the dry weight of the arsenic, a sulfonate wetting agent (Tenlo 400) reduced dustiness of the arsenic so that it could be applied with a spreader with less hazard to the operator. Pelleted arsenic was less toxic than powder in greenhouse trials during the first two croppings. By the third it was equal to the powder. Tests on combinations of arsenic tri- oxide and CMU prove that they are not antagonistic, that their toxic effects are additive, and that the combination is effective the first year. Plot trials Plot trials with arsenic trioxide con- firmed the lack of toxicity during the year of application. They emphasized the fact that where perennial weeds are present an additional soil sterilant of greater solubility and teachability should be incorporated. On railroad ballast pelleted arsenic was more effective than the powder. The pellets were not moved by the wind from passing trains, and they tended to gravitate into crevices where weed seeds germinate. Results on a series of plots at various elevations along the Southern Pacific line through the Sierra Nevada moun- tains correlated with annual precipita- tion better than with any other variable. High precipitation brings about more rapid action by the chemical. Arsenic trioxide applications have proved effective for controlling weeds along the lining of irrigation ditches, around cattle guards, bridgeheads and similar structures, and paved airstrips. Field-scale use The largest use of arsenic trioxide has been on fire lines in the mountain regions. These include firebreak systems and cleared strips along roads, railroads, and powerlines. Machine application is practical on smooth terrain ; hand appli- cation has been required in rough, steep, and rocky sites. Arsenic sterilization on sandy or gravelly soils in steep terrain may result in gully erosion. In such situations pro- vision for adequate drainage should be made. All arsenic-sterilized strips should be posted to avoid disturbance by grading machinery. Such strips provide con- venient routes for travel by such equip- [27 ment, and often volunteer smoothing 0.1 per cent by weight (1 pound per breaks the arsenic cover and allows re- 1,000 pounds) of a wetting agent and 0.2 infestation. per cent of water. The wetting agent and the water should be thoroughly mixed, Powder form or pellets then added to the arsenic a bit at a time The dust or powder form is satisfac- with thorough mixing, tory if wind action is not excessive. Arsenic can be pelleted by using a The operator making the application lignin binder and screening to retain should protect his hands and face from particles between 10 and 35 mesh. The accumulation of the arsenic dust. binder should be non-hygroscopic and Dustiness of the powdered form of ar- sufficiently water soluble that the pellets senic can be eliminated by incorporating melt readily when moistened. Tests have not been conducted on the palatibility and acceptability of arsenic trioxide to animals when wetting agents and pelleting binders are included. Anyone using arsenic with such materials added should run an acceptability test before apply- ing in areas accessible to animals. LITERATURE CITED Crafts, A. S. 1935a. Factors influencing the effectiveness of sodium chlorate as a herbicide. Hilgardia 9: 437-58. 19356. The toxicity of sodium arsenite and sodium chlorate in four California soils. Hilgardia 9:461-98. 1935c. Plot tests with sodium arsenite and sodium chlorate as soil sterilants in California. Calif. State Dept. Agr. Mo. Bui. 24 (4, 5, 6) : 247-59. 1939a. Toxicity studies with chlorate in eighty California soils. Hilgardia 12:233-47. 19396. The relation of nutrients to toxicity of arsenic, borax, and chlorate in soils. Jour. Agr. Res. 58:637-71. Crafts, A. S., H. D. Bruce, and R. N. Raynor 1941. Plot tests with chemical soil sterilants in California. California Agr. Expt. Sta. Bui. 648: 1-25. Crafts, A. S., and C. W. Cleary 1936. The toxicity of arsenic, borax, and chlorate, and their combinations in three California soils. Hilgardia 10:401-13. Crafts, A. S., and R. N. Raynor 1936. The herbicidal properties of boron compounds. Hilgardia 10: 343-71. Crafts, A. S., and R. S. Rosenfels 1939. Toxicity studies with arsenic in eighty California soils. Hilgardia 12:177-200. Robbins, W. W., A. S. Crafts, and R. N. Raynor 1942. Weed control. A textbook and manual. 1st ed. McGraw-Hill Book Co., New York. 543 p. Rosenfels, R. S., and A. S. Crafts 1939. Arsenic fixation in relation to the sterilization of soils with sodium arsenite. Hilgardia 12:203-09. ACKNOWLEDGMENTS Credit is due James C. Harvey, who set up the greenhouse experiments and estab- lished many of the field plots reported here. Harold R. Drever completed the green- house testing. 8Jm-2,'54(A8634)BEB