730 



SCIENCE. 



[N. S. Vol. XI. No. 280. 



rigating stream away from where it seems 

 to go. The winds may blow down and de- 

 stroy the branches and their fruit, and 

 frosts sometimes bring an untimely end to 

 a vigorous stand of plants. Thieves may 

 break through and steal. 



Weather we cannot control, but long- 

 season crops should be started early. The 

 enemies of the street need to be fenced out 

 as securely as possible, but it is not easy to 

 put a pad-lock upon a field experiment. 

 Poison for the vermin, insecticides for the 

 insects and fungicides for the diseases 

 should be used with a judicious hand. One 

 needs to guard his experiment area with in- 

 finitely more care than an ordinary orchard, 

 farm or garden crop, for the fruits are an 

 in-gathering of truth, as nearly as it can 

 be obtained. 



The Keeping of the Record. — This is the 

 most difi&cult of all the operations connected 

 with the successful issue of a field experi- 

 ment. The conscientious superintendent 

 will not trust to his memory for the details, 

 nor write up the results from the informa- 

 tion that is only in mind at the close of the 

 experiment. He must keep a record of the 

 stages of the crop from the time the ground 

 is broken until the end of the harvest. 

 This means during the more rapidly grow- 

 ing season the taking of almost daily notes 

 as to growth of plants, etc., associated with 

 rainfall and soil and air temperatures. 

 The character of the notes will vary with 

 the crop, but they, in any case, need to be 

 full. 



Many plans of note-taking have been 

 devised, but of them all there is nothing 

 better than a day-book for all the crops, 

 from which the items are transferred upon 

 rainy days and odd hours to the separate 

 books for each particular crop. There is a 

 journal, so to speak, for the beans, or each 

 class of beans, as bush, dwarf lima and pole 

 sorts, those of the peas, potatoes, tomatoes 

 and cucumbers, etc. A set of small pigeon- 



holes placed upon the wall above the desk 

 holds the crop -books close at hand. ISTote 

 paper in pads has been tested, but the loose 

 pages, while convenient for [some things, 

 are easily lost. Form-sheets with spaces 

 for certain entries, and upon set dates with 

 various schedules prove cumbersome, and 

 from all the methods the writer has ac- 

 cepted the books as the most satisfactory. 



Statement of Results. — The statement of re- 

 sults of plot experiments should of course 

 be clear and comprehensive. The precise 

 point to be held up to view will vary with 

 the crop and the reason for the trial. In 

 general it should be in percentage of in- 

 crease, or of decrease, over the control plot 

 where the single feature of the test did not 

 obtain. For example, if the experiment is 

 with a remedy for the pod spot of the beans 

 the results admit of several methods of ex- 

 pression. They may be in terms of total 

 weight of plants, of weight of marketable 

 pods, or matured sound seed, or in percent- 

 age of diseases as compared with the con- 

 trol area. It seems to the writer that the 

 chief point is the relative amount of market- 

 able pods of the check plot. In other words, 

 place in concise form the increase of the 

 salable product, for this is the crucial test 

 of the value of the fungicides. To this may 

 be added the cost of the gain that the 

 practicability of the spraying may be shown. 

 As secondary matters the effect upon the 

 whole plant and sundry other matters may 

 be given. 



But the end is not yet ; to return to the 

 increase of salable product. Shall it be 

 shown in pounds or quarts or in number of 

 pods. Snap beans are sold by the quart 

 and it might seem the most natural method 

 to state it in terms of quarts. In getting 

 the results upon a plot it is easier to weigh 

 the pods, and to assort out the spotted from 

 the diseased pods it is simpler to count 

 them. All these methods have been tested 

 and experience suggests the counting as the 



