1. Sixteen colchicine treatments, involving five main methods of application and six concentrations, have been used on Triticum interspecific crosses, Triticum—Aegilops intergeneric crosses and Agropyron—Triticum intergeneric crosses.2. The efficacy of the major treatments in terms of plant survival, plant fertility and ear fertility has been compared, and the most successful method of application has been found to be absorption through the cut leaves—i.e. capping cut-back tillers with a small glass phial containing colchicine.3. Fertility induced by colchicine is not complete. In plants producing grain, every ear is not necessarily fertile, nor are all spikelets in fertile ears usually fertile.4. Evidence has been obtained of differential cross response to colchicine action, and also of cross-group response to particular treatments.

No abstract

1. In general the juvenile growth of interspecific Triticum and intergeneric Tritioum-Aegilops amphidiploids, was normal and did not deviate from the parental behaviour. However, in the case of the interspecific hexaploids derived from T. monococcum and T. aegilopoides there was an early check in growth, and all the amphidiploids derived from A. caudata developed more or less severe chlorosis.2. Most amphidiploids were intermediate between their parents in mature plant characters. However, in the hexaploid Triticum interspecific amphidiploids there was a transgressive increase in plant height, and whilst the derivatives of Triticum x Aegilops crosses fell within the range of their parents, they were closer to one or other parent depending upon the Aegilops species involved.3. The ear morphology of the Triticum interspecific amphidiploid was generally a composite of that of the parents, the influence of each parent being detectable. In the Triticum-Aegilops amphidiploids, however, ear morphology was so strongly influenced by the Aegilops species, that there was frequently no obvious difference between the ears of amphidiploids with a common Aegilops parent but different wheat parents.4. The fertility of the amphidiploids was low and there was no significant correlation between the fertility of amphidiploids and parents. There was a significant correlation between both the grain weight, and number of spikelets per ear, in parents and amphidiploids, but whilst the mean grain weight was about that of the parent with the heavier grain, the mean of spikelets per ear was about that of the parent with the fewer spikelets.

No abstract

1. Chromosome pairing has been studied in twenty-two different sterile F1 hybrids involving the genera Aegilops, Agropyron and Triticum, together with their colchicine derived amphidiploids having chromosome numbers of 2n = 42, 56 and 70. Cytological evidence has been correlated with male and female fertility, while chromosome pairing in the parents has been studied in relation to their amphidiploids.2. Some of the sterile F1 hybrids showed little or no pairing, while in others the pairing was appreciable. There was an association of the amount of pairing with the parental combinations used in the production of the hybrids in that the interspecific hybrids were characterized by a relatively high degree of pairing, particularly those with 28 chromosomes, while the intergeneric hybrids either lacked pairing or showed a low incidence.3. In the A1 amphidiploid generation, chromosome pairing was in all cases high, and in some cases almost complete. In all cases multivalent formation in the amphidiploid was lower than bivalent formation in its undoubted F1 hybrid. Different amphidiploids showed various degrees of differential affinity. Univalent formation occurred in some amphidiploids, while bivalent formation in some was increased by a loss of chromosomes.4. In all cases there was a reduction in chiasmata per nucleus and chiasmata per bivalent in the amphidiploid compared with its parent species. Reduction values were not directly associated with any increase in chromosome number of the amphidiploid, nor with the presence of multivalents.5. No confirmation could be obtained of the view that multivalent formation in amphidiploids is a more generally sensitive index of chromosome homology than bivalent formation in the undoubted F1 hybrid. The absence of multivalents in an amphidiploid does not disprove the existence of structural chromosome homologies between the two parents.