404 



NATURE 



[September 21, igu 



especially to take more hopeful views of their absolute 

 purification than have hitherto prevailed. It is at present 

 accepted as part of the natural perversity of things that 

 most high-class seed crops must throw " rogues," or that 

 at the best the elimination of these waste plants can only 

 be attained by great labour extended over a vast period of 

 time. Conceivably that view is correct, but no one 

 acquainted with modern genetic science can believe it 

 without most cogent proof. Far more probably we should 

 regard these rogues either as the product of a few definite 

 individuals in the crop, or even as chance impurities brought 

 in by accidental mixture. In either case they can pre- 

 sumably be got rid of. I may even go further and express 

 a doubt whether that degeneration which is vaguely sup- 

 posed to be attendant on all seed crops is a physiological 

 reality. Degeneration may perhaps affect plants like the 

 potato which are continually multiplied asexually, though 

 the fact has never been proved satisfactorily. Moreover, it 

 is not in question that races of plants taken into unsuitable 

 climates do degenerate rapidly from uncertain causes, but 

 that is quite another matter. 



The first question is to determine whether a given rogue 

 has in it any factor which is dominant to the corresponding 

 character in the typical plants of the crop. If it has, then 

 we may feel considerable confidence that these rogues 

 have been introduced by accidental mixture. The only 

 alternative, indeed, is cross-fertilisation with some distinct 

 variety possessing the dominant, or crossing within the 

 limits of the typical plants themselves occurring in such a 

 way that complementary factors have been brought together. 

 This last is a comparatively infrequent phenomenon, and 

 need not be considered till more probable hypotheses have 

 been disposed of. If the rogues are first crosses the fact 

 can be immediately proved by sowing their seeds, for 

 segregation will then be evident. For example, a truly 

 round seed is occasionally, though very rarelv, found oil 

 varieties of pea which have wrinkled seeds. I have three 

 times seen such seeds on my own plants. A few more were 

 kindly given me by Mr. Arthur Sutton, and I have also 

 received a few from M. Philippe de Vilmorin— to both of 

 whom I am indebted for most helpful assistance and advice. 

 Of these abnormal or unexpected seeds some died without 

 germinating, but all which did germinate in due course 

 produced the normal mixture of round and wrinkled, 

 proving that a cross had occurred. Cross-fertilisation in 

 culinary peas is excessively rare, but it is certainlv some- 

 times effected, doubtless by the leaf-cutter bee (Megachile) 

 or a humble-bee visiting flowers in which for some reason 

 the pollen has been inoperative. But in peas crossing is 

 assuredly not the source of the ordinary rogues. These 

 plants have a very peculiar conformation, being tall and 

 straggling, with long internodes, small leaves, and small 

 flowers, which together give them a curious wild look. 

 When one compares them with the tvpical cultivated plants 

 which have a more luxuriant habit, it seems difficult to 

 suppose that the rogue can really be recessive to such a 

 type. True, we cannot say definitely a priori that any one 

 character is dominant to another, but old preconceptions are 

 so strong that without actual evidence we always incline 

 to think of the wilder and more primitive characteristics 

 as dominants. Nevertheless, from such observations as I 

 have been able to make, I cannot find any valid reason for 

 doubting that the rogues are reallv recessives to the type. 

 One feature in particular is quite inconsistent with the 

 belief that these rogues are in any proper sense degenerative 

 returns to a wild type, for in several examples the rogues 

 have pointed pods like the cultivated sorts from which they 

 presumably been derived. All the more primitive 

 kinds have the dominant stump-ended pod. If the rogues 

 had the stump pods they would fall into the class* of 

 dominants, but they have no single quality which can be 

 declared to be certainly dominant to the type, and I see no 

 reason why they may not be actually recessives to it after 

 all. Whether this is the true account or not we shall know 

 for certain next year. Mr. Sutton has given me a quantity 

 of material which we are now investigating at the lohn 

 Innes Horticultural Institution, and bv sowing the seed of 

 a great number of individual plants separately T anticipate 

 that we shall prove the rogue-throwers to be a class a 

 The pure types then separately saved should, according to 

 expectation, remain rogue-free, unless further sporting or 

 NO. 2l86, VOL. 87] 



fresh contamination occurs. If it prove that the long and 

 attenuated rogues are really recessive to the shorter and 

 more robust type, the case will be one of much physiological 

 significance, but I believe a parallel already exists in the 

 case of wheats, for among certain crosses bred by Prof. 

 Biffen, some curious spelt-like plants occurred among the 

 derivatives from such robust wheats as Rivet and Red Fife. 



There is another large and important class of cases to 

 which similar considerations apply. I refer to the bolting 

 or running to seed of crops grown as biennials, especially 

 root crops. It has hitherto been universally supposed that 

 the loss due to this cause, amounting in Sugar Beet as it 

 frequently does to five, or even more, per cent., is not 

 preventable. This may prove to be the truth, but I think 

 it is not impossible that the bolters can be wholly, or 

 almost wholly, eliminated by the application of proper 

 breeding methods. In this particular example I know that 

 season and conditions of cultivation count for a good deal 

 in promoting or checking the tendency to run to seed; 

 nevertheless one can scarcely witness the sharp distinction 

 between the annual and biennial forms without suspecting 

 that genetic composition is largely responsible. If it proves 

 to be so, we shall have another remarkable illustration of 

 the direct applicability of knowledge gained from a purely 

 academic source. " Let not him that girdeth on his hai 

 boast himself as he that putteth it off," and I am quite 

 alive to the many obstacles which may lie between the 

 conception of an idea and its realisation. One thing, 

 however, is certain, that we have now the power to 

 formulate rightly the question which the breeder is to put 

 to nature ; and this power and the whole apparatus bv 

 which he can obtain an answer to his question — in what- 

 ever sense that answer may be given — has been derived 

 from experiments designed with the immediate object of 

 investigating that scholastic and seemingly barren problem, 

 " What is a species? " If Mendel's eight years' work had 

 been done in an agricultural school supported by public 

 money, I can imagine much shaking of heads on the 

 County Council governing that institution, and yet it is 

 no longer in dispute that he provided the one bit of solid 

 discovery upon which all breeding practice will henceforth 

 be based. 



Everywhere the same need for accurate knowledge is 

 apparent. I suppose horse-breeding is an art which has bv 

 the application of common sense and great experience been 

 carried to about as high a point of perfection as any. 

 Yet even here I have seen a mistake made which is obvious 

 to anyone accustomed to analytical breeding. Among a 

 number of stallions provided at great expense to improve 

 the breed of horses in a certain district was one which was 

 shown me as something of a curiosity. This particular 

 animal had been bred by one of the provided stallions out 

 of an indifferent country mare. It had been kept as an 

 unusually good-looking colt, and was now travelling the 

 country as a breeding stallion, under the highest auspices. 

 I thought to myself that if such a practice is sanctioned 

 bv breeding acumen and common sense. Science is not, 

 after all. so very ambitious if she aspires to do rather 

 better. The breeder has continually to remind himself that 

 it is not what the animal or plant looks that matters, hut 

 what it is. Analysis has taugfit us to realise, first, that 

 each animal and plant is a double structure, and next that 

 the appearance may show only half its composition. 



With respect to the inheritance of many physiological 

 qualities of divers kinds we have made at least a beginning 

 of knowledge, but there is one class of phenomena as yet 

 almost untouched. This is the miscellaneous group of 

 attributes which are usually measured in terms of size, 

 fertility, yield, and the like. This group of characters has 

 more than common significance to the practical man. 

 Analysis of them ran nevertheless only become possible 

 when pure science has progressed far beyond the point vet 

 reached. 



I know few lines of pure research more attractive and at 

 the same time more likely to lead to economic results than 

 an investigation of the nature of variation in size of the 

 whole organism or of its parts. By what factors is it 

 caused? By what steps does it proceed? By what limita- 

 tions is it Beset? In illustration of the application of these 

 questions I may refer to a variety of topics that have been 

 lately brought to my notice. In the case of merino sheep 



