268 



differ for different insects, and perhaps the insecticides will even have a 

 different ratio. Therefore the standard for experiment must be made 

 arbitrarily and calculated to the comparison standard for each insect. 

 The past season we have made the experiment standard for our work 

 approximately the amount of arsenic in the insecticides. Next season 

 we intend to use the same standard, but first have the substance ana- 

 lyzed and make our mixtures as accurately as possible. In preparing 

 for a set of experiments we have invariably made but one strength of 

 the poison and obtained the other strengths by dilution, both for the 

 sake of speed and accuracy. 



Not only must the strength of the mixture be brought to a standard, 

 but the method of its i:)reparation must be uniform. The addition of 

 any substance, as lime, may entirely change the nature of the mixture, 

 and if water from different sources be used, it must be proven by experi- 

 ment that no variation in the results is produced. The length of time 

 after the x^reparation is made has a great influence. As tested on sor- 

 _ghum leaves, London purple is but little if any changed by standing, 

 while Paris green becomes fully twice as injurious and the injury from 

 white arsenic is increased perhaps a hundred fold. The Paris green 

 mixture used on the oak, even when fresh, is more injurious than Lon- 

 don purple, a fact hard to account for. 



There is so much difference in the action of the poison on different 

 leaves that it is necessary to carefully select leaves that are uniform 

 for experimental purposes. I believe that individual leaves are most 

 desirable for several reasons. Among the things that are proven to 

 cause variation in the results are, age, physical condition, and exposure. 

 Uniformity in these conditions at least must be secured when experi- 

 ments are to be compared. 



For marking the leaves, I have tried a number of methods. Tags at- 

 tached by fine wire are quite quickly applied, but injure the petioles of 

 some i)lants. Tying is a rather tedious operation, but without other 

 objections. A faster way is the use of artist's oil paints, putting a mark 

 or number on the leaf. I think that a set of conductor's punches would 

 make the quickest and most satisfactory mark. 



A most important thing is the reading of the results of the experi- 

 ments. Figures are for all reasons better than the use of such terms as 

 much and little. My plan is to read in tenths. I find I can do this 

 very uniformly. It is my habit to occasionally test myself by rereading 

 sets of experiments. With a higher number, as twenty, I can not read 

 satisfactorily. Ten is a most convenient number for calculations. 



It is best to take more than one reading, perhaps as often as daily, 

 and to trace the injury carefullj-. This will enable one to note excep- 

 tional leaves and perhaps discover the cause. The difference observ- 

 able between early and late reading suggests the possibility of arresting 

 injury by a thorough washing with water, or better, by applying water 

 containing iron rust, and would not the use of the latter remove the 

 objections to the use of the arsenic on fruits ? 



