May 11, 1900.] 



SCIENCE. 



731 



best method. It is less subject to error in 

 the harvest, there is no diflSculty with frac- 

 tions, the unit is a natural one, and shows 

 upon the face of the results that each 

 pod has been inspected. Neither weight 

 nor measure do full justice to the disease 

 for the spotted pods average much smaller 

 than the healthy ones. Finally from the 

 numbers the percentages are made up and 

 the gain or loss is instantly applied to what- 

 ever unit of quantity that is in vogue. 



The results must be calculated to a full 

 stand of plants for both the trial and check 

 plots. The absurdity of anything otherwise 

 is evident with such crops as tomatoes or 

 eggplants, but it is equally rational with 

 bush beans or turnips. Sow the seed thickly 

 and thin the crop to a given distance in the 

 row, say six inches ; count the plants at 

 harvest for each row and if there are any 

 missing allow for them. 



Let the following serve as an illustration : 

 A plot of ground is limed at the rate of 300 

 bushels per acre for the club-root in turnips. 

 The adjoining plot has everything in com- 

 mon excepting the lime. The plants stand 

 six inches apart and those upon the limed 

 land flourish to the harvest time. Those 

 upon the check plot languish and die from 

 week to week and at the time the record is 

 made only one-third are to be found. The 

 roots are pulled and weighed, the roots 

 weighed separately, assorted into those with 

 and those without club-roob and each lot 

 counted and weighed. From the counts the 

 percentages are made up, while the weights 

 also go on record. One large unclubbed 

 root will outweigh a dozen clubbed ones and 

 the unfairness of weights as a basis for the 

 final per cents, is apparent. 



With the writer the results of his field 

 experiments square usually to the line of 

 disease, and the number of units of the 

 plant product is the one that best conforms 

 to the requirement of the case. AVith vege- 

 table fruits as cucumbers, tomatoes, pep- 



pers, eggplants, it is the fruit and whether 

 it is marketable or diseased. Tests have 

 been made with the fruits left upon the 

 plants until a single harvest day, as with 

 cucumbers. The results may then become 

 very striking to the eye ; but the writer 

 thinks the better way is to pick at the same 

 intervals as for market and keep the record 

 in the book. The results of the spraying 

 may here be perhaps better estimated by 

 counting the healthy and the diseased leaves 

 and getting the percentage from these. The 

 gain in marketable product is also given. 



In root crops the writer has not always 

 followed the plan herein advocated. Thus 

 with sweet potatoes the series of experiments 

 for checking the soil rot, started out five 

 years ago with the pound as the unit and 

 the plan has not been changed. It is not as 

 easy and probably is not as accurate, owing 

 to the fact that diseased roots are on the 

 average smaller than healthy ones, because 

 they are diseased. 



In like manner in estimating the scab 

 upon the round potato very scabby potatoes 

 kept in a glass jar for the purpose and taken 

 as one hundred per cent., have served for 

 estimating the amount of disease of each 

 pile at harvest time regardless of size or 

 number, the crop being recorded in total 

 yield in pounds and percentage of scab. 

 For example weight of tubers 75 pounds, 

 per cent, of scab 60 ; weight 84 pounds, per- 

 centage of scab 15. 



The value of pictures is not to be over- 

 looked as when the relative yields are placed 

 in baskets, boxes or heaps. The superiority 

 of the one treatment, whether of fertilizer, 

 culture, pruning or spi'aying over another, 

 in this way appeals to the eye in a vivid 

 manner . Squares or circles upon the printed 

 page will aid in the same way. There is, 

 taken all in all, perhaps no better method of 

 indicating the relative results than by the 

 parallel bars, and they admit of many feat- 

 ures in vei-y small space. Thus three or 



