226 



NATURE 



[October 14, 1920 



he \ 

 }le 1 



be ' 



brought in one from each side of the pedigree instead 

 of both being contributed by one parent, we get a^.'iin 

 a result in which the two parental combinations 

 occur more frequently, the two recombinations or 

 "cross-overs " less often than we should expect. In 

 the first case the two characters appear to hang 

 together in descent to a certain extent, but not com- 

 pletely; in the latter, similarly to repel each other. 



This type of relationship has been found to be of 

 very general occurrence. The linked characters do 

 not otherwise appear to be connected in any way 

 that we can trace, and we therefore conclude that 

 the explanation must be sought in the mechanism of 

 distribution. Two main theories having this funda- 

 mental principle as their basis, but otherwise distinct, 

 have been put forward, and are usually referred to as 

 the reduplication and the chromosome view respec- 

 tively. The reduplication view, proposed by Bateson 

 and Punnett (Proc. Roy. Soc, 1911), rests on the 

 idea that segregation of factors need not necessarily 

 occur simultaneously at a particular cell-division. 

 The number of divisions following the segregation 

 of some factors being assumed to be greater than 

 those occurring in the case of others, there would 

 naturally result a larger number of gametes carrying 

 some factorial combinations and fewer carrying 

 others. If this differential process is conceived as 

 occurring in an orderly manner it would enable us 

 to account for the facts observed. It has, however, 

 to be said that we cannot say why segregation should 

 be successive, nor at what moments, on this view, 

 it must be presumed to occur. On the other hand, 

 the conceptions embodied in the chromosome hypo- 

 thesis as formulated by Morgan and his fellow- 

 workers ("The Mechanism of .Mendelian Heredity" 

 (Morgan, Sturtevant, MuUer, and Bridges), 1915) are, 

 in these respects, quite precise. They are built 

 around one cardinal event in the life-cycle of animals 

 and plants (some of the lowest forms excepted), viz. 

 the peculiar tvpe of cell-division at which the number 

 of chromosomes is reduced to half that to be found 

 during the period of the life-cycle extending back- 

 wards from this moment to the previous act of 

 fertilisation. In the large number of cases already 

 investigated the chromosome number has been found, 

 as a rule, to be the same at each division of the 

 somatic cells. We can, in fact, take it as estab- 

 lished that it is ordinarily constant for the species. 



These observations lend strong support to the view 

 that the chromosomes are persistent structures — that 

 is to say, that the chromatin tangle of the resting 

 nucleus, whether actually composed of one continuous 

 thread or not, becomes resolved into separate chromo- 

 somes at corresponding loci at each successive 

 mitosis. The reduction from the diploid to the 

 haploid number, according to the more generally 

 accepted interpretation of the appearances during the 

 meiotic phase, is due to the adhering together in 

 pairs of homologous chromosomes, each member of 

 the set originally received from one parent Iving 

 alongside and in close contact with its mate received 

 from the other. Later, these bivalent chromosomes 

 are resolved into their components so that the two 

 groups destined one for either pole consist of whole 

 dissimilar chromosomes, which then proceed to divide 

 again longitudinallv to furnish equivalent half 

 chromosomes to each of the daughter nuclei. 



The obvious close parallel between the behaviour of 

 the chromosomes — their pairing and separation — and 

 that of Mendelian allelomorphs which similarly .show 

 pairing and .segregation, first led to the suggestion 

 that the factors controlling somatic characters are 

 located in the.se structures. The ingenious extension 

 of this view which has been elaborated bv Morgan 



NO. 2659, VOL. 106] 



and his co-workers presumes the arrangement of the 

 factors in linear series after the manner of the visible 

 chromomeres — the beadlike elements which can 

 seen in many organisms to compose the chromatin 

 structure — each factor and its opposite occupying cor- 

 responding loci in homologous chromosomes. From 

 this conception follows the important corollary of the 

 segregation of the factors during the process of 

 formation and subsequent resolution of the bivalent 

 chromosomes formed at the reduction division. We 

 should suppose, according to Morgan, in the case of 

 characters showing independent inheritance and giving 

 identical Mendelian ratios whichever way the mating 

 is made and however the factorial combination is 

 brought about, that the factors controlling the several 

 characters are located in different chromosomes. 

 Thus in the case of Datura already mentioned the 

 two factors affecting sap-colour and prickliness respec- 

 tively would be presumed to be located so far apart 

 in the resting chromatin thread that when separation 

 into chromosomes takes place thev become distributed 

 to different members. Where unrelated characters 

 show a linked inheritance the factors concerned are 

 held, on the other hand, to lie so near together that 

 they are always located in one and the same chromo- 

 some. Furthermore — and here we come to the most 

 debatable of the assumptions in Morgan's theory — when 

 the bivalent chromosome composed of a maternal and 

 a paternal component gives rise at the reduction divi- 

 sion to two single dissimilar chromosomes, these new 

 chromosomes do not always represent the original in- 

 tact maternal and paternal components. It has been 

 observed in many forms that the bivalent structure 

 has the appearance of a twisted double thread. 



Already in 1909 cytological study of the salamander 

 had led Janssen (La Cellule, xxv.) to conclude 

 that fusion might take place at the crossing points, 

 so that when the twin members ultimately draw- 

 apart each is composed of alternate portions of the 

 original pair. Morgan explains the breeding results 

 obtained with Drosophila by a somewhat similar hypo- 

 thesis. He also conceives that in the process of 

 separation of the twin lengths of chromatin cleavage 

 between the two is not always clean, portions of the 

 one becoming interchanged with corresponding seg- 

 ments of the other, so that both daughter chromo- 

 somes are made up of complementary sections of the 

 maternal and paternal members of the duplex 

 chromosome. On the main issue, however, both 

 schemes are in accord. .-X physical basis for the 

 phenomenon of linlcage is found in the presumed 

 nature and behaviour of the chromosomes, viz. their 

 colloidal consistency, their adhesion after ]ilh-ing at 

 the points of contact when in the twisted condition, 

 and their consequent failure to separate cleanly before 

 undergoing the succeeding division. 



-According to Morgan, the frequency of separation 

 of linked characters is a measure of the distance apart 

 in the chromosome of the loci for the factors con- 

 cerned, and it becomes possible to map their position 

 in the chromosome relatively to one another. In this 

 attempt to find in cytological happenings a basis for 

 the observed facts of inheritance, our conception of 

 the material unit in the sorting-out process has been 

 pushed beyond the germ-cell, and even the entire 

 chromosome, to the component sections and particles 

 of the latter structure. 



To substantiate the " chromosome " view the 

 primary requisite was to obtain proof that a particular 

 character is associated with a particular chromosome. 

 With this object in view it was sought to discovVr a 

 type in which individual cTiromosomes could be 

 identified. Several observers working on different 

 animals found that a particular chromosome differing 



