1919] HORTICULTURE. 43 



production Of a large number of cherries on a short length of branch. The trees 

 are small and the indications are that to make this variety profitable two or three 

 times as many trees should be set per acre as of the typical Arabian coffee. 



In a fertilizer experiment conducted with 40 plats of 3 young coffee trees each, 

 the first crop at three years from seed showed a large increase in yield from 

 the plats which have received nrtrogen. Of the 12 plats which gave the highest 

 production nitrogen had been used in the fertilizer applied to 10 of them. In a 

 fertilizer test with older trees, in which applications of a complete chemical 

 fertilizer and stable manure had been made twice annually for a number of 

 years the fertilized plat produced 73.4 per cent more than the check, whereas in 

 the preceding season their yields were nearly uniform. 



In a lime and nitrogen test with coffee trees grown for two years in 5-gal. 

 cans, 18 plants grown in limed soil differed from the 3 check plants in average 

 weight by only a very small fraction of 1 per cent. The lime was applied in a 

 series of amounts ranging in rate from 0.5 ton to 16 tons per acre. In the 

 division which had received nitrogen there had been made at intervals of six 

 months three applications of 8 gm. each per tree of ammonium sulphate to one 

 group and to the others sodium nitrate ranging in amount from 4 to 1G gm. per 

 tree per application. The weight of the trees fertilized with ammonium sulphate 

 averaged 22.1 per cent greater than the check, and those fertilized with sodium 

 nitrate averaged 22.9 per cent greater than the check. The trees which had 

 received 4, 8, and 10 gm. of sodium nitrate fell below, while those which had 

 received 1- and 16 gm. of sodium nitrate surpassed, those which had received 

 8 gm. of ammonium sulphate, 16 gm. of sodium nitrate producing an increased 

 weight of 50.3 per cent over the check. The production of coffee cherries 

 averaged 87 per cent greater for the nitrogen-fertilized trees than for the check. 



The work with cacao consisted chiefly in the collection of data as to the 

 product of individual trees. From the older orchard set in 1903, the yield for 

 the calendar year 1916 exceeded that of any preceding year, though more than 

 one-fifth of the trees produced nothing. 



Cultural experiments with vanilla were continued. In a planting test with 

 48 tip cuttings of 10 nodes each, half were left to wilt in a fairly well-shaded 

 place for an interval of 12 days between cutting and planting, the others being 

 set immediately. Half were planted in a mixture of equal parts of clay and 

 river sand and hall' in leaf mold. Cuttings which had been wilted for 12 days 

 before planting gave equally good root development when grown for the same 

 length of time as cuttings planted without wilting. Root development was 85 

 per cent greater in leaf mold than in soil. Since both the wilted and unwilted 

 cuttings gave vigorous and highly satisfactory growth, it is suggested that the 

 planter's convenience should determine which method to follow. 



Blossoms were observed to open in the spring on vine growth made as late as 

 the preceding autumn. A period of approximately two months was observed to 

 elapse from the pushing out of the inflorescence bud and the opening of the first 

 blossom. After blossoming the pods attained full length in six to eight weeks, 

 though seven to nine months were required for maturing. Fruit set from more 

 than 90 per cent of the hand-pollinated blossoms in a recorded series and from 

 1.5 per cent of the blossoms in a series not hand-pollinated. Working steadily 

 the author pollinated in one hour 237 blossoms, indicating that hand-pollination 

 need not be costly as to time required for the operation. 



In curing small quantities of vanilla, blistering was found to result from 

 sweating the pods between scalding and drying. Sometimes 20 to 50 per cent 

 of the pods developed watery blisters. The omission of this sweating reduced 

 the proportion of blistered pods to less than 2 per cent, but an extract expert 

 reported the quality of the unsweated pods as inferior to that of the sweated 



