HOETICULTURE. 535 



produced only 18 per cent of seed stalks. Sandy soil gave 30 per cent of seed 

 stalks in the case of early planting and but a slight percentage in the case of 

 late planting. The results as a whole seem to indicate that early planting, soil 

 overrich In organic matter, or soil too abundantly supplied with sand will 

 encourage premature seeding. 



A comparison of ridge and level cultivation in growing onions resulted in 

 favor of the former method. Ridging apparently induced early maturity and 

 the bulbs seemed to " bottom up " better. 



Report of the horticulturist, C. F. Kinman {Porto Rico Sta. Rpt. lOl-ff, 

 pp. 17-22, pi. 1). — ^A short progress report on investigations with fruits, vege- 

 tables, and miscellaneous plants (E. S. R., 31, p. 634). 



Among the East Indian varieties of mangoes being tested Amini and Cam- 

 bodiana were the most promising during 1914. Over 400 trees, including GO 

 imported varieties, were under cultivation. Work with yautias, dasheens, 

 yams, sweet potatoes, etc.. is being continued. Cooperative variety tests have 

 been conducted near San Juan on the sandy coastal plains. The first crop of 

 yautias grown on the sandy soils has been unsatisfactory, but few of the 

 varieties proving suflBciently prolific to warrant commercial cultivation. 

 Dasheens were more promising, though not heavy producers. Some varieties of 

 sweet potatoes have given good returns, while many are not profitable under 

 these conditions. 



Some data are given on the cooperative fertilizer experiments with citrus 

 fruits, the results of which have been published in full in a bulletin by the 

 station (E. S. R., 33, p. 241). The cooperative experiments with coconuts have 

 been extended to parts of the island where different weather conditions pre- 

 vail. Eight harvests have been made since the fertilizer experiments were 

 started in 1912. The trees under observation are old palms on sandy beach 

 land suitable for coconut culture, but have been slow to respond to fertilizer 

 treatment. In the last harvest, in plats given a complete fertilizer, the yield 

 was in excess of that in check plats, although only small differences existed 

 between the complete fertilizer and incomplete fertilizer and check plats. 

 Data collected on individual coconut trees have shown a wide variation in 

 yield, indicating the importance of seed-nut selection by those contemplating 

 a new grove. 



Experiments are being conducted to determine the advisability of growing 

 leguminous cover crops in a young coconut grove. Jack beans (Canavalm 

 cnsifonnis) and a few species of velvet beans (StizoloMinn spp. ) planted in 

 April all made good growth and crowded out all native weeds and grasses. 

 The Stizolohium spp. appeared to have a longer growing season and to produce 

 a heavier crop of vegetation than the jack beans and are therefore preferable 

 in coconut plantations. Leguminous cover-crop work was also conducted in 

 citrus orchards and pineapple plantations. Much of the work was with the 

 velvet bean species. These crops have made an excellent growth in both hea\-y 

 and light soils where rainfall has been suflBcient. They demand frequent cut- 

 ting back, however, to keep them from injuring the trees. Notes are given 

 on different species of velvet beans desirable for cover-crop planting. Both 

 jack beans and pigeon peas were found to be injurious to the two varieties 

 of pineapples under observation, Cabezona and Red Spanish, when planted in 

 beds with them. The pigeon pea caused the most injury and the Cabezona 

 pineapple was more severely injured than the Red Spanish. 



A number of varieties of pineapples were planted more than two years ago 

 in an area between 7-year-old PithecoIoMum saman trees to test the effect of 

 the shade and the nitrogen stored in the soil by the leguminous trees on the 

 development of the pineapples. The shade trees were planted 20 by 20 feet 



