I04 Journal of Agricultural Research voi. xvii, no. 3 



tion does not prevent great differences in the crop from year to year. 

 Some growers hold that there is a good fruit crop only during seasons 

 ynih. favorable weather for bees at blooming time. Hedrick (7) analyzed 

 the weather records of New York with respect to fruit production and 

 showed that in general unfavorable weather is the dominant factor in 

 crop failures. In fact, for a long time fruit growers have recognized 

 certain weather combinations as detrimental to or prohibiting the 

 setting of fruit. 



If weather is to be assigned such an important role in relation to fruit 

 fulness, the question arises as to the significance of the great variation 

 in the time of bloom from year to year. For instance, plums have 

 varied nearly one month in the time of flowering at the Fruit- Breeding 

 Farm in the last seven years, the earliest bloom in this period beginning 

 April 24, 1915, and the latest May 20, 1916. The cause for such a varia- 

 tion in time of bloom should not be assigned entirely to the weather 

 of early spring, because Sandsten {13) found, upon analyzing the bloom- 

 ing records at Madison, Wis., that the time of flowering was influenced 

 more by the growing conditions of the preceding summer and fall than 

 by those of the spring. In Plate 15 the prevailing weather of early 

 spring when plums are in flower is presented in some detail. It will be 

 seen from the analysis presented in these graphs that cool weather and 

 frequent rains can be expected in Minnesota for a period of even greater 

 length than that covered by the greatest extremes in the time of bloom. 

 Therefore, inasmuch as a range in blooming time of as much as one 

 month has not meant an escape from periods of unfavorable weather, 

 early or late blooming does not necessarily have a constant relation to 

 fruitfulness. 



The period of 10 days after bloom was selected (PI. 15) because it covers 

 for the most part the time of fertilization. In only 10 instances out of 

 142 did the minimum temperature occur in the day and the maximum 

 at night, so that the curve of maximum temperature may be considered 

 as the day temperature and that of the minimum as the night tempera- 

 ture. In the graph for each season the period of bloom is indicated by 

 the lighter shaded portion between the maximum and minimum temper- 

 ature curves. In the case of wind and the character of the day (sun- 

 shine or cloudiness) a 12-hour day was taken because of the bearing of 

 wind and sunshine on bee flight. The date in the graph is located in the 

 midpoint, which is 12 m. The short, broken-line curves indicate the 

 wind velocity during the daytime only, i. e., from 6 a. m. to 6 p. m. 

 The legend is at the right of the graph. The character of the day is 

 ' shown by the shading at the base of each graph ; the dark bar represents 

 the portion of the day which was cloudy, the cross bar that which was 

 partly cloudy, and the white the time of sunshine. A dotted line is 

 drawn through each graph at the 40° and 51° F. points, the former 



