and showed no injury from poisons, except in the case of one or two trees, and even here very slight. Two or three months afterwards, however, or toward the end of July, many of the trees treated showed a varying percentage, not large, but yet noticeable, of the leaves which were turning yellow and falling to the ground, apparently going through the same stages of ripening and falling which is characteristic of the autumnal foliage. The effect of arsenic on the plant, therefore, seems to hasten maturity of foliage, and perhaps also of fruit. That this is prejudicial in some degree to the plant can hardly be questioned, and must be weighed against the destruction of foliage or fruit by insects to determine the advisability of poisoning.

Relative value of poisons against insects.-A large series of feeding tests were made with the arsenicals referred to above, half-grown fall webworms being used in the experiments throughout. From 20 to 50 larvæ were caged for each test and fed on leaves wetted with freshly prepared poisons. In all a total of 1,057 larvæ were used in the experiments. Fresh food was poisoned and supplied to the larvæ every day or two during the time of the experiments. The series of experiments were three times repeated in the case of Paris green, arsenite of copper, and London purple, and only once or twice performed with the other arsenicals. Against these larvæ the poisons were used at the strength of 1 pound to 160 gallons, 1 pound to 100 gallons, and, with the excep tion of London purple, 1 pound to 80 gallons. One pound of poison to 160 gallons was rather unsatisfactory, although of considerable destructive effect against the larva, causing the death within nine days of an average of 71 per cent of the larva with Paris green and arsenite of copper and 20 per cent with London purple. The rate of 1 pound to 100 gallons was eminently successful with all the poisons except London purple, which seemed to be inferior to the others and to vary greatly in the different samples. Very little difference was shown between the arsenicals with and without lime, and as between ordinary Paris green, pulverized Paris green, and arsenite of copper there is little to choose in the matter of effectiveness. Arsenite of lead, which I believe has not hitherto been used as an insecticide, gave an excellent record and seemed to be, if anything, more rapid in its action than the other arsenicals.

In the tests at the rate of 1 pound to 80 gallons of water all the larvæ were killed within seven days, no essential variation being shown between the four poisons used, viz, Paris green (ordinary and pulverized) and arsenite of copper and arsenite of lead, both with and without lime.

The following table includes the later tests at the strength of 1 pound to 100 gallons of water, and is introduced to illustrate the details of the result. Except with London purple, the larvae were all killed. within nine days, and for the most part had succumbed within six days. The slower and inferior action of the London purple is clearly shown,

Arsenicals.-1 pound to 100 gallons of water, with and without lime.

Much of the variation in the length of time necessary to effect the death of larvæ is explained by the condition of the larvæ. If they are just molting or about to molt, a day or two will be lost, and it may happen that when they are put on poisoned food, not relishing it, they refuse to eat before molting, so that two or three days may intervene before they feed at all heartily on the poisoned food. In the case of these experiments it was very evident that the larva detected almost immediately the fact of the unnatural condition of the foliage, refusing to eat or eating very scantily, whereas larvæ placed at the same time on unpoisoned food fed with great freedom and greediness. These results explain the apparently slow action of poisons on larvæ which every experimenter has witnessed. It simply means that the larvæ dislike the poisoned leaves and refuse to feed until forced to by extreme 5850-No. 6—3

hunger. In the meantime, however, the protection to the foliage is just as great as though the larvæ were killed outright.

Confirmatory of the efficacy of the arsenite of lead are some experiments conducted on the grounds of the Department against the bagworm on Arbor vitæ. Some half dozen trees thickly infested with these larvæ were sprayed at the rate of 1 pound to 50 gallons of water. Ten days after the application the cases of the bagworms were invariably found to contain dead larvæ, and the cessation of feeding of these larvæ was almost immediate after the application. The use of these arsenicals on pear and cherry trees was also accompanied by a demonstration of value as protective applications, in that trees not included in the experiment were much eaten by the pear slug.

Notes on the composition and other characteristics of the arsenicals.—A great many analyses of the old arsenicals, Paris green and London purple, have been made and are valuable for comparison. Paris green is a definite chemical compound and should have a nearly definite composition. London purple is a waste product and naturally shows considerable variation in composition. Dr. Lintner quotes an analysis showing 43.65 per cent of arsenious oxide in this insecticide, whereas other analyses have shown it to contain as little as 19.64 of arsenious oxide. Scheele's green, theoretically, should contain 52.94 per cent arsenious oxide, and the lead arsenites and arsenates occur in several different forms and have a variable component of arsenious oxide. The common form of lead arsenite contains 47.03 per cent; the arsenate such as used by the gypsy moth commission 28.53 per cent, and the lead arsenite used in my own experiments shows by analysis 39.34 per cent of arsenious oxide. In other words, Paris green and Scheele's green are not widely different in the active principle. Lead arsenite in its most poisonous form approaches somewhat closely these two, whereas the arsenate is very much weaker and yet costs more to manufacture. The following solution tests of these arsenicals have been kindly made for me by Prof. George Patrick, of the Division of Chemistry:

The samples were mixed with Potomac water at the rate of 1 pound of the powder to 100 gallons of water, kept in contact for twenty-four hours, with frequent shakings during the first ten hours. The quantities of arsenic dissolved are indicated by the figures given above.

The solubility of the poison depends, as might have been expected, on the fineness of the powder, as indicated by Paris green, ordinary and pulverized. The greater amount of soluble arsenic in No. 3 is not more in the same proportion relative to its greater fineness than between the two samples of Paris green. The great solubility of London purple is shown, and also the slight solubility of the arsenites of lead. The conditions of the experiment represent an exaggeration of what might occur in spraying operations.

Paris green, ordinary, is a rather coarse crystal, and settles very rapidly in water solution. London purple is pulverulent, and settles slowly. The lead arsenites and arsenates are precipitates, very pulverulent, and settle much more slowly than even London purple, and in this; category also comes arsenite of copper, which is quite as satisfactory in this respect as the lead poisons. The attempt to pulverize Paris: green was only partially successful. The powder was noticeably finer than the ordinary product, but still settled very rapidly in water. The: fineness of division of the poison, as indicated in the foregoing experiment, has an effect also in making it more poisonous to plants, but this: may be overcome by the use of lime, and it will probably prove that a less amount of the more finely divided poison will accomplish as much as a greater amount of coarser powder.

The cost of these arsenicals varies with the market price of the constituent elements, as bluestone, arsenic, lead, etc. The price of Paris green ranges pretty close to 20 cents per pound. Arsenite of copper

or Scheele's green can be made for 10 cents per pound, and the arsenite of lead (pink arsenite) prepared for our use can be made for about 8 cents per pound.

Conclusion.--Results of experiments with these arsenicals during the past few years have demonstrated to the satisfaction of the writer the value of the simple arsenite of copper, and the experiments of the present year also show the very considerable value of the arsenite of lead. Both have a very decided advantage in their more fine division, which render their use much more safe than an insecticide which rapidly settles to the bottom of the tank, resulting in likelihood of great inequality in strength as sprayed on the trees. Both of these arsenicals may be obtained at about one-half cost of Paris green, or approximately the cost of London purple, and are both of them far superior to London purple in effectiveness and in safety when applied to foliage.

Half the cost of Paris green or aceto-arsenite of copper, as has been previously pointed out, comes in making it more suitable for a pigment, its original market use, and this adds nothing to its value against insects. I think the time has come when we should abandon its use as an insecticide and substitute for it Scheele's green. This arsenical will be manufactured in any quantity by the Adler Color and Chemical Works, of New York City, and as there is no secret about its composition it can also be made by any firm working in such material.

Arsenite of lead is a most promising insecticide. The only objection to it now is the fact that it may be made in several different forms hav ing varying percentages of arsenious oxide. Further work with this substance is now being conducted, which, it is hoped, will demonstrate a method of producing it even more cheaply than that now employed and at the same time give a guaranty of uniformity in composition. It may be colored, as in the sample exhibited, to prevent its being mistaken for harmless substances.

Mr. Howard read the following:

ON THE FUTILITY OF TRUNK AND CROWN WASHING AGAINST THE ELM LEAF-BEETLE.

By L. O. HOWARD, Washington, D. C.

When the elm leaf-beetle made its first serious outbreak in New Haven, Conn., in the early summer of 1895 the authorities were unprepared and the seriousness of the matter did not enforce itself until the damage was practically done. It was then too late to spray, and, in fact, no apparatus was ready. They immediately, however, as the writer stated before this association last year, began extensive work with kerosene emulsion upon the larva on the lower part of the trunk and about the base of the tree. Nearly 14,000 elms were treated in this way by the superintendent of parks, and many barrels of standard kerosene emulsion were made under his supervision and distributed to citizens who came for it with buckets to use upon trees in their private grounds. It was considered that the scale upon which this work was done was so great that it would exercise a very considerable influence upon the number of the beetles during the summer of 1896. That it did exercise a certain amount of such influence can not be doubted, but that the insect can be kept in check by such means is entirely negatived by the number of insects which successfully hibernated and laid their eggs upon leaves the present season. In the discussion of Dr. Lintner's very interesting paper upon "The elm leaf-beetle in Albany" at our last meeting Dr. Lintner stated in regard to the published accounts that many of the larvæ had dropped from the tree instead of traveling down the trunk, although the idea seemed reasonable it was not confirmed by his observations. He instanced a Scotch elm upon his own grounds which has long and somewhat drooping branches which extend over an extension of the house, but he was not able to find at any time during the season a single larva or pupa upon the roof of the extension, although fallen leaves in corners offered convenient resting places for them. Nor had he noticed in all of the hours that he passed under the elms in his study of the insect one of the larvæ upon his clothing or known it to occur upon others. Referring to this statement, the writer in the course of his discussion remarked that in his opinion not above 60 per cent of the insects upon large trees crawl down the trunk to the ground, the others falling from the branches direct or undergoing the transformations upon the rough bark or upon larger trees in the main crotches. In this statement the writer was corroborated by Mr. Smith, who further stated that he had known them to climb fences and transform in the crevices.

The statement made at the last meeting that not above 60 per cent of the insects crawl down the trunks of large trees under normal conditions, although it seemed to the writer to indicate an estimate which was about correct, must be altered, as the result of certain observations