The ftsZ gene of Escherichia coli, which codes for an essential cell division protein, is subjected to multiple regulation, as shown in part with studies using an ftsZ::lacZ operon fusion located on phage λJFLIOO. Using this same fusion, we sought to isolate regulatory mutants overexpressing ftsZ by selecting mutants able to grow on lactose. One Lac+ mutant was obtained which overexpressed the ftsZ::lacZ fusion 70-fold. The mutation responsible for the overexpression lies in a new gene, cot, located near 56 min on the E. coli genetic map. The cot mutation probably affects the transcription of a chromosomal open reading frame, 0RF1, lying downstream of the bioA gene and adjacent to the ftzZ::lacZ fusion of the λJFL100 prophage integrated at attλ. Using an ftsZ84(Ts) strain, in which there was a double selection for overexpression of both ftsZ::lacZ and ftsZ+, no Lac+Tr mutants were obtained from 3·6 × 1010 bacteria; the introduction of a mutL allele, increasing spontaneous base substitution mutation rates 75-fold, did not permit us to isolate such a mutant. We conclude that Lac+ftsZ-constitutive mutations cannot be obtained in λJFL100 lysogens by a single base substitution.

Data are presented concerning the gene arrangements in the second and third chromosomes of Drosophila robusta in eight altitudinal transects. A consistent change is the increase in the arrangement 2L-3, particularly in the linkage combination 2L-3.2R, with increasing altitude. The reciprocal decrease with increasing altitude affects several different 2-left arrangements, most consistently 2L-1. The arrangements of 2-right show no significant variation with altitude, and those of 3-right do so only in a few samples of the two northern transects studied, none in any of the southern ones. These results confirm previous evidence for the significant role of the arrangements of the left arm of the second chromosome in the adaptations of this species to altitude and suggest further that interactions of linked arrangements are involved in these adaptations. The data also indicate that the factors responsible for the altitudinal adaptations of this species are in many cases not the same ones that are responsible for variations in its gene arrangements with latitude.

Simple, inexpensive techniques were used to analyse the mtDNA of nine chromosomally distinct populations of Petrogale. Eight of these populations occur in sequence along the Great Dividing Range of eastern Australia; six have been described as species. Diagnostic mtDNA morphs were found throughout the latitudinal ranges of four of the described species. A fifth morph spanned the ranges of two described species and three additional taxa which have been designated chromosome races. These five mtDNA morphs, and others with local distributions, were used to assess interactions between the taxa. Limited introgression was indicated across the chromosomal boundaries of P. penicillata/P. herberti and P. inornata/P. assimilis; atypical mtDNA morphs were found within the P. herberti and P. inornata chromosomal distributions. No introgression was detected between P. herberti and P. inornata, whose distributions are separated by the Fitzroy River. Nor was there evidence of recent contact between P. assimilis, P. herberti and P. purpureicollis, despite the late occupancy of parts of the intervening area by unidentified Petrogale. These data, considered in the light of information obtained from previous studies on chromosomes, allozymes and parasites, have contributed to the decision to consider all the eastern representatives of the lateralis-penicillata group of Petrogale as being specifically distinct from each other.

We have carried out an investigation of the effects of various parameters on the accumulation of deleterious mutant alleles in finite diploid populations. Two different processes contribute to mutation accumulation. In random-mating populations of very small size and with tight linkage, fixation of mutant alleles occurs at a high rate, but decreases with extremely tight linkage. With very restricted recombination, the numbers of low-frequency mutant alleles per genome in randommating populations increase over time independently of fixation (Muller's ratchet). Increased population size affects the ratchet less than the fixation process, and the decline in population fitness is dominated by the ratchet in populations of size greater than about 100, especially with high mutation rates. The effects of differences in the selection parameters (strength of selection, dominance coefficient), of multiplicative versus synergistic selection, and of different amounts of inbreeding, are complex, but can be interpreted in terms of opposing effects of selection on individual loci and associations between loci. Stronger selection slows the accumulation of mutations, though a faster decline in mean fitness sometimes results. Increasing dominance tends to have a similar effect to greater strength of selection. High inbreeding slows the ratchet, because the increased homozygous expression of mutant alleles in inbred populations has effects similar to stronger selection, and because with inbreeding there is a higher initial frequency of the least loaded class. Fixation of deleterious mutations is accelerated in highly inbred populations. Even with inbreeding, sexual populations larger than 100 will probably rarely experience mutation accumulation to the point that their survival is endangered because neither fixation nor the ratchet has effects of the magnitude seen in asexual populations. The effects of breeding system and rate of recombination on the rate of molecular evolution by the fixation of slightly deleterious alleles are discussed.

Wright proposed that there is a ‘shifting balance’ between selection within demes, random drift, and selection between demes at different ‘adaptive peaks’. We investigate the establishment and spread of new adaptive peaks, considering a chromosome rearrangement, and a polygenic character under disruptive selection. When the number of migrants (Nm) is small, demes fluctuate independently, with a bias towards the fitter peak. When Nm is large, the whole population can move to one of two stable equilibria, and so can be trapped near the lower peak. These two regimes are separated by a sharp transition at a critical Nm of order 1. Just below this critical point, adaptation is most efficient, since the shifting balance greatly increases the proportion of demes that reach the global optimum. This is so even if one peak is only slightly fitter than the other (ΔW≈1/N), and for both strong and weak selection (Ns (Ns ≪ 1 or Ns ≫ 1). Provided that Nm varies sufficiently gradually from place to place, the fitter peak can be established in regions where Nm≈1, and can then spread through the rest of the range. Our analysis confirms Wright's argument that if selection, migration and drift are of the same order, the ‘shifting balance’ leads to efficient evolution towards the global optimum.