308 



SCIENCE. 



[N. S. Vol. XI. No. 269. 



physical conditions of the marshes are dis- 

 cussed, the flora described and the segrega- 

 tion of the vegetation into formations illus- 

 trated. The physiological and structural 

 peculiarities of the leading plants of each 

 formation are then described. Finally an 

 attempt is made to formulate the results 

 into a series of principles of general appli- 

 cation. 



Complications in Citrus hybridization caused by 

 polyembryony : Me. H. J. Webbee, De- 

 partment of Agriculture. 

 During the last five years the writer has 

 made about a thousand hybrids of the vari- 

 ous species of Citrus commonly cultivated, 

 and has had under his care and observa- 

 tion about an equal number made by Mr. 

 W. T. Swingle. Last spring a careful com- 

 parison of the foliage characters of these 

 hybrids led to the interesting observation 

 that where two or three seedlings were de- 

 veloped from a single seed, as was fre- 

 quently the case, they not infrequently 

 showed marked foliage differences. The 

 fact that the common orange and many 

 other species of the genus Citrus are com- 

 monly polyembryonic is well known. A 

 single seed of the common orange has been 

 known to produce as high as thirteen dif- 

 ferent seedlings, although it is seldom that 

 more than three of the embryos are capable 

 of development. Strasburger, in his critical 

 study of the polyembryony of this group, 

 found that the embryos, other than that 

 developed from the fecundated egg cell, are 

 derived from certain cells of the nucellus 

 lying near the embryo sac wall which be- 

 come specialized, grow and divide rapidly, 

 and form a tissue mass which pushes 

 out into the embryo sac and forms an 

 embryo similar to that formed in the nor- 

 mal way from the egg cell. The embrj'os 

 formed in this way Strasburger called 

 adventive embi-yos. If we correctly un- 

 derstand the action of fertilization it is 



clear that in hybridizing fruits of this sort 

 only those embryos developed from the egg 

 cell proper, as a result of the fecundation, 

 would show an indication of the hy- 

 bridization. The adventive embryos de- 

 veloped directly from the mother tissue we 

 should not expect to show any of the char- 

 acters of the male parent. This was the 

 conclusion reached by Mr. Swingle and the 

 writer jointly in discussing the matter sev- 

 eral years ago, and the development of the 

 hybrids has now shown this to be the case. 

 In several instances of hybrids of Citrus 

 aurantium, which has unifoliolate leaves, 

 with C. trifoliata, which has trifoliolate 

 leaves, where the former was used as the 

 female parent, two and three seedlings have 

 been produced from the same seed, one of 

 which had trifoliolate leaves and the others 

 strictly unifoliolate leaves exactly like those 

 of the mother parent. In such cases it 

 is evident that the trifoliolate seedling 

 inherits this character from the male par- 

 ent and that the embryo from which it 

 grew was developed from the egg cell proper. 

 The other seedlings in such cases, having 

 unifoliolate leaves like the mother parent, 

 are doubtless developed from the so-called 

 adventive embryos. The writer has also 

 observed the same phenomenon in the re- 

 ciprocal hybrids, Citrus trifoliata $ X C 

 aurantium S , and in the hybrids of Citrus 

 nobilis $ X C. aurantium $ . The observa- 

 tions have been sufficiently extended so that 

 we may be certain of the common occur- 

 rence of the phenomenon in citrous hybridi- 

 zation. 



In hybridizing citrous fruits to secure im- 

 proved sorts this fact unfortunately intro- 

 duces a serious complication. The ma- 

 jority of citrous hybrids, in almost every 

 case resemble the female parent in foliage 

 characters, and it will thus be seen that 

 until they fruit it will be impossible to de- 

 termine whether they are true hybrids, 

 showing a preponderating influence of the 



