The chromosome types of three species,Hyacinthus orientalis,Bellevallia romana andScilla nutans, have been examined at various stages; the inter-relation of tlie complements does not point to any affinity between the species or to any common descent from a three-chromosome ancestor. The complement of the Hyacinth is not capable of a purely tetraploid interpretation because one of the four long chromosomes has a definite and constant second constriction while the other three are alike in every respect.

The condition of the constrictions in the three species was followed at different stages: certain constrictions are permanent, others appear at one stage but may be suppressed at another. The phenomenon seems to be the result of a local discontinuity which occurs in different chromosomes at different positions—bisecting the chromosome or cutting off a minute element, which will form, in the contracted state, a “satellite.” Reasons have been shown for believing that the responsiveness in the chromosome to the attraction of the poles is more diffuse than is usually imagined and that this zone of response need not coincide with a constriction.

The appearance of the constriction in all cases may be accounted for by supposing that two chromosome elements are joined by a fine single, or double, chromatin thread surrounded, no doubt, by the film-coat that envelopes the whole chromosome.

Secondary splitting of the chromosomes in the first microspore division was noticed at all stages inHyacinthus orientalis, and inScilla nutans in prophase.

Counts have been made of chromosome numbers in the pollen grains of triploid Hyacinths: various theoretical explanations of the proportions in which they occur have been tested and the results have been found to agree closely with the assumption that each of the extra chromosomes undergoes a constant chance of loss, of the order of one-tenth. The occurrence of somatic variations in chromosome number was noted in heteroploid varieties ofHyacinthus orientalis and inScilla nutans.

The pollen mother-cell divisions in triploid tulips are described. At prophase only pairs of homologous chromosomes associate at any particular point, but different pairs associate at different points. The fact that three threads lie side by side proves that each corresponds to a whole chromosome rather than to half a chromosome; the parasynaptic interpretation is thus substantiated.

It is suggested that the two distinct stages of prophase are, first, that in which attraction exists between four threads, and, secondly, that in which attraction exists between only two of these threads.

A new type of variegation, inVicia Faba, is shown by a heterozygote for which the corresponding homozygotes are albino and green. The expression of this factor in both pure and hybrid state is shown to be affected by five other conditions, some of which may in effect counterbalance the differences between the genetical factors.

Local differences in the breeding behaviour of the plant are attributed to (i) maternal influence and (ii) failure of embryos of various genetical constitutions under unsuitable maternal conditions.

The inheritance of the dwarf, blotchless and fawn-blotched characters is described.

Fritillaria Meleagris has twelve pairs of chromosomes, two with median, ten with more or less snb-terminal attachment constrictions. One pair has a second constriction.

The basic chromosome number inPyrus is seventeen. Cultivated varieties are all orthoploid. Aneuploid seedlings are poor and abnormal.

The sweet cherries (Prunus avium), which were previously reported to have extra chromosomes, are found from root tips to be diploid (2x=16). They cannot therefore be derived directly from hybridisation with sour cherries, as hitherto believed.

1. The determination of the relationship of crossing-over to chiasmata and of chiasmata to segregation has made it possible to define the conditions of variation of pairing chromosomes in complex-heterozygotes and sex-heterozygotes and to analyse their behaviour in genetic terms. The object of the present studies is, first, to test the predictions made with regard to sex-chromosomes deducing their genetic structure from cytological observations and, secondly, to note those respects in which the behaviour of the sex-chromosomes throws light on the problems of chromosome mechanics.

2. The sex-chromosomes of the Norway rat each consist of a pairing segment and a differential non-pairing segment. The pairing segment includes the spindle attachment and chiasmata may be formed on one or both sides of it so that the first division is either reductional or equational for the differential segments (Text-fig. 17).

3. The differential segments will have complete sex-linkage, the pairing segments will have partial sex-linkage diminishing in proportion to the crossing-over distance from the differential segments.

4. The shape, movements, and staining capacity of the sex-chromosomes in the rat and elsewhere agree in suggesting that they have a lower surface charge than the autosomes, and this is held to be responsible for the special mechanism ofX-chromosome segregation in organisms lacking theY.

The simplest method of origin of ring formation, direct segmental interchange, gives interchange heterozygotes which inevitably contain segments of chromosomes of three different types in respect of crossing-over;differential segments with no crossing-over,interstitial segments (proximal to the interchange) with reduced and deleterious crossing-over, andterminal segments with simple crossing-over. These different kinds of segments account for the existence of a complex, of genes characteristically associated with the complex, and of others in which freer crossing-over is possible and no characteristic association with the complex occurs.

This type of reproduction is an example of a compromise between the advantages of free combination and of absolute linkage found in simple sexual reproduction and in clonal reproduction respectively. This compromise is found very widely in plants and animals, although expressed in many different ways. Wherever it occurs it is a means of species formation.

1. Fragments pair with major chromosomes inTradescantia by lateral chiasmata whose chromatid structure is revealed by pre-treatment. They do not then develop the usual major spiral.

2. Less contracted chromosomes in mutant cells have more coils. Spiralization therefore consists in reducing the number of coils.

3. Interlocked chromosomes, like multivalents, may have linear, twisted, or discordant orientations according to the method of interlocking.

4. The type of distribution of the bivalents on the metaphase plate inT. bracteata depends on the numbers of their chiasmata and on the radial or tangential adjustment of the individual bivalents.

5. The central chromosomes of the metaphase plate are pushed to the edge of the group during anaphase in some forms and species. This change may take place asymmetrically both at mitosis and meiosis. It shows that the repulsion from the poles is still acting on the chromosomes while they are moving towards the poles. Thus the anaphase in some forms comes to correspond with the hollow metaphase in others.

6.Rhoeo discolor and diploid, triploid and tetraploid species ofTradescantia show the results of inversion crossing-over under normal conditions. Fragments are small inRhoeo and in the tetraploids, showing that the assumption of pairing and crossing-over being restricted to the end regions is justified.

Following misdivision of the centromere at meiosis in diploid and triploidFritillaria new telocentric chromosomes are formed whose broken ends rejoin within the centromere. This type of chromosome is delayed at metaphase and anaphase in the pollen-grain mitosis. It may then, either break again at the centromere, or pass without separation to the pole as a new iso-chromosome.

The misdivision and the origin of the iso-chromosome are each likely to be important as affecting the genetic structure of the chromosome and the mechanical properties of the centromere.

1. The frequency of chiasmata and of univalents is similar in pollen and embryo sac mother cells ofLilium testaceum.

2. Cells with fewest chiasmata have those chiasmata most strongly localized either proximally or distally.

3. Hence pairing must have begun either near the centromere or near an end. The contact point is optional.

4.M chromosomes come into contact more readily near the ends,S chromosomes near the centromere.

5. ButM chromosomes pair more slowly, so that the frequency of their chiasmata falls away more rapidly in low chiasma cells than that ofS’s.

6. Samples of cells used for the study of meiosis in species and crosses must therefore be regarded as representing cross-sections of the process of pairing secured by a variable interruption of the process. Statistical treatment can be used to indicate the order of pairing and the means of interruption.

Irradiation affects the time scale of development by which we measure its effects in the breakage of chromosomes. This explains why single-time comparisons have hitherto shown discordant results. Our comparative breakage-time series enable us to measure both effects and to show direct proportionality of dose and breakage in the pollen grain.

Comparison of pollen grains and pollen tubes shows that their chromosomes have a similar breakability and also inherent rejoinability (as deduced from the frequencies of minutes). But actual rejoining, as between different chromosomes or chromosome arms, such as is revealed in the empirical coefficient of reunion, is frustrated by the close packing of the nucleus.

The pollen-tube nuclei show extreme overlapping of B′ and B″ owing to extreme divergence in rates of nuclear development. They also show a greater primary effect with greater stickiness than in the pollen and (perhaps correlated with this) SR of unbroken ends, which overlaps with the regular SR of broken ends. The general result agrees with that inferred from progeny tests after X-rayingDrosophila sperm while of course showing many other changes that would not appear in any progeny.