PLANT BREEDING AND HYBEIDIZATION. 215 



being- found in a panicle containing many thousand tlowcr.s. In the 

 author's -experiments the whole panicle was sown in l)oxes in the hope 

 of the presence of some fertile seed. On account of the minuteness 

 of the tlowers and the rarity of the production of fertile seed, the 

 ordinary method of cross fertilization could not l)e adopted. Stami- 

 nate plants were planted to the windward of the pistillate ones, or 

 in alternate rows. In another series the panicles were covered with 

 bags and later dusted with pollen-bearing- plants of known value. It 

 was found that frequently the pollen was infertile, while the pistillate 

 Howers were fertile. The most valuable varieties so far obtained 

 through seed canes have followed the principal characteristics of the 

 stand nate parent. 



So far as the writer's oV)servation has gone, the seed canes tiller 

 more extensivel}" than the plants from the cane top. The Ribbon and 

 White Transparent Aai'ieties have proved the best for mother plants, 

 the mother plants governing the size, color, and to a considerable 

 extent the sugar content of the cane. Purity of the juice has not 

 been definitely determined as due to either parent. As to the yield 

 of sugar, the average for the island of Barbados is said to be 1.7 tons 

 per acre, while one variety shown by the writer, designated as No. 208, 

 yields 3 tons of sugar per acre at Barbados, and its value has l)een 

 further attested on the islands of Trinidad, St, Kitts, etc. 



A paper on the Cytological Aspects of Hj'brids, by W. A. Cannon, 

 of Columl)ia University, New York, showed that the relation between 

 the cytological and experimental studies could not l)e definitely stated 

 in the present state of our knowledge on the subjects. A review was 

 given of Mendel's laws in the light of modern cytological studies. So 

 far these studies have been made on first-generation hyl^rids. It is 

 said that the normal division of sex nuclei leads to fertilit}^ in 

 hybrids, but abnormal division to sterility. Cytological studies on 

 cotton and other hybrids, it is claimed, show (1) a possible cause of 

 sterility, (2) that variation in the hybrid may or may not be associated 

 with variation in spermatogenesis, and (3) that chromosomes tend to 

 retain their respective individualities, as shown in many hybrids 

 examined. 



Improvement of Roses by Bud Selection was the subject of a paper 

 by L. C. Corbett, of this Department, in which experiments were 

 r(»ported which were undertaken to determine the relative value of 

 blind and flowering wood in rose production. It was shown that 

 individual characteristics of a branch were perpetuated from genera- 

 tion to generation in plants asexually propagated, and also that cumu- 

 lative results are not to be expected by the selection of parts showing 

 like tendencies through successive generations. The flowering habit 

 of plants produced from flowering wood through five generations was 

 in no way increased, nor was it diminished when blind wood was 



