March 20, 1903.] 



SCIENCE. 



451 



btit in practice we shall probably find 

 many more exceptions to the formulse than 

 confirmations of them, even though the ex- 

 ceptions can be explained, after we find 

 them, by Mendel's principle of heredity. 

 It has been said that we shall soon be 

 able, as a result of Mendel's discoveries, 

 to predict varieties in plant-breeding. Be- 

 fore considering this question, we must 

 recall the fact that a cultural variety is a 

 succession of plants that have characters 

 sufficiently marked and uniform to make 

 it worth cultivating in place of some older 

 variety. Now and then it may be worth 

 while to introduce some new energy or new 

 trend into a general lot of offspring by 

 making wholesale crosses, not expecting 

 ever to segregate any particular variety 

 or strain from the progeny ; but these eases 

 are rare, and the gain is indefinite and 

 temporary. So far as our knowledge at 

 present goes, I see no warrant for the hope 

 that we can predict varieties with any de- 

 gree of exactness, at least not beyond a 

 very narrow effort. Following are some of 

 the reasons that seem to me to argue against 

 the probability of useful prophecy of vari- 

 ties, so far as the Mendelian results are 

 concerned: (1) We do not know what 

 plants will Mendelize until we try. (2) 

 Even in plants that do Mendelize, only 

 half of the offspring have stable charac- 

 ters. But we can not predict for even this 

 half, for it is impossible to determine be- 

 forehand which seeds showing dominant 

 characters (and these are three fourths of 

 the offspring) will 'come true.' Domi- 

 nance, as we have seen, is of two kinds in 

 respect to its behavior in the next genera- 

 tion—constant and hybrid ; and the hybrid 

 dominance, which is twice as frequent as 

 the other, breaks up into constant domi- 

 nance, hybrid dominance and recessiveness. 

 (3) Mendel's law deals primarily with mere 

 characters, not with a variety or with a 

 plant as a whole. Every plant is a com- 



posite of a thousand characters, and from 

 the plant-breeder's point of view there may 

 be as many undesirable characters as de- 

 sirable ones. No plant is perfect; if it 

 were there would be no need of plant- 

 breeding. The breeder wants to preserve 

 the desirable characters or traits and elim- 

 inate the undesirable ones, but under the 

 strict interpretation of Mendelism this is 

 difficult. The one germ gamete and the 

 one sperm gamete that unite to make the 

 new plant each contain all the alternative 

 characters; these characters are bound to 

 reappear in the offspring, and all that the 

 breeder gains is a new combination or ar- 

 rangement of characters, and undesir- 

 able attributes may be as troublesome as 

 before. (4) The breeder usually wants 

 wholly new characters as well as recom- 

 binations of old ones, or he wants aug- 

 mented characters. For example, a car- 

 nation grower wants a four-inch flower, but 

 he has only three-inch flowers to work with, 

 and augmentation of characters is no part 

 of the original Mendelian law. Perhaps 

 these augmented and new characters are 

 to be got by means of ordinary variation 

 and selection, or other extra-crossing 

 means; but we know, as a matter of fact, 

 that augmented characters do sometimes 

 appear in hybrids. (5) New and unpre- 

 dictable characters are likely to arise from 

 the influence of environment or other 

 causes, and these may be recorded in the 

 gametes and vitiate the final results. (6) 

 Variability itself may be a unit character, 

 and therefore pass over. There is prob- 

 ably such a thing as a 'tendency to vary,' 

 wholly aside from the fact of variation. 

 (7) Many of the plants with which we 

 need most to work in plant-breeding are 

 themselves eminently variable, and the re- 

 sults, even if there is true Mendelism, may 

 be so uncertain as to be wholly unpredict- 

 able. (8) Many plants with which we 

 must work: will not close-fertilize. Some 



