June 2, 1911] 



SCIENCE 



869 



explain the more rapid injury from samples 

 containing unhydrolized pyro-arsenate. 



The Ammonia Test. — The reaction between 

 ammonia water and the acid arsenates of lead 

 may be used to test the presence of these com- 

 pounds in a sample. 



A fair-sized sample (10 to 20 grams) is 

 worked up in water (25 to 50 c.c.) and an 

 equal volume of strong ammonia (26° B.) is 

 added. This is digested with heat, agitated 

 and finally brought to the boiling point. The 

 sample is then allowed to cool and settle, and 

 the clear solution decanted through a filter. 

 The filtrate is boiled until the ammonia is 

 nearly or all driven off. The remaining solu- 

 tion is then made distinctly acid with acetic 

 acid, and a concentrated solution of lead ace- 

 tate containing free acetic acid is added. 

 Any arsenic acid which the filtrate may carry 

 will then be precipitated in the form of acid 

 arsenate of lead. With no precipitate form- 

 ing at first, add the lead acetate solution to 

 large excess and allow to stand. A preciptate 

 may appear in a few minutes or an hour. 



With this treatment pyro-arsenate, hydro- 

 gen-arsenate and mixtures of ortho and pyro 

 give abundant white precipitates. In any 

 case where the precipitate forms at once and 

 can be designated as more than a trace, i. e., 

 renders the filtrate opaque, the sample will 

 prove injurious under foliage test conditions, 

 and will probably prove injurious in commer- 

 cial spraying operations, especially where cli- 

 matic conditions favor such injury. On the 

 other hand, samples which show no ammonia 

 test are practically free from foliage-injuring 

 properties. 



The ammonia reaction is complete and may 

 be used for the quantitative determination of 

 any arsenic oxide that may be present in ex- 

 cess of that required to form the ortho-com- 

 pound. 



Errors arising from Chemical Tests and 

 Analysis. — It is not enough to know that the 

 ratio of As^Oj to PbO is as 1 to 2.90 or more. 

 The essential thing is whether all the PbO is 

 combined with the As^O^. Chlorine and or- 

 ganic bodies may retain sufficient lead to 

 materially affect the nature of the compound. 



In case the arsenic oxide content is slightly 

 greater than it should be, and an estimable 

 quantity of chlorine is present, the sample 

 will certainly give an ammonia test and prove 

 injurious to foliage under test conditions. 



The true ortho samples will probably show 

 an excess of lead oxide amounting to 1 to 

 4 per cent. 



The ammonia test is positive except in 

 rare cases where the uncombined lead is suffi- 

 cient to take up the liberated arsenic oxide. 



Water Solubility. — This subject has re- 

 ceived much attention from chemists but as 

 usually handled, does not indicate the true 

 condition of the sample. Most arsenicals 

 yield only a limited portion of their arsenic 

 to a given amount of water, but may repeat 

 this a very great number of times. As has 

 already been shown, the liberation of arsenic 

 oxide from the acid arsenates depends on 

 transposition to ortho under neutral and al- 

 kalin conditions. However, the amount of 

 arsenic acid taken up by a given volume of 

 water can not exceed a very small quantity 

 before the reaction of the solvent becomes acid 

 and stops the transposition. As a matter of 

 fact, such solutions are neutral or alkalin to 

 litmus. 



The water solubility of arsenate of lead is 

 readily shown in a qualitative way by digest- 

 ing in water for a time (ten to twenty-four 

 hours), decanting through a filter, acidifying 

 the filtrate with acetic acid and then adding 

 lead acetate solution as in the ammonia test. 

 A white precipitate or turbidity indicates 

 water-soluble arsenic oxide. This operation 

 should be repeated a number of times to deter- 

 mine whether the amount of dissolved arsenic 

 oxide remains a constant. With this method 

 of treatment ortho- and acid-arsenate samples 

 show a decided difference in behavior. Both 

 arsenates may show about the same arsenic 

 oxide content per unit volume of water for 

 the first two or three washings. After that, 

 however, the quantity of arsenic removed 

 from the ortho sample will be very much 

 reduced, but that from the acid sample will 

 remain practically constant for a large num- 

 ber of washings. A device by which a con- 



