I investigate models of the spread of transposable elements, such as the Drosophila melanogaster P elements, that can exist in autonomous and non-autonomous forms. Elements which have their major impact on host fitness in the process of transposition can, under certain conditions, come to a stable balance between transposition and selection. This stable balance for autonomous elements can be disrupted by the invasion of further elements, which do not produce a transposase enzyme, and may produce a repressor of transposition. I examine this secondary invasion process, and show that a stable equilibrium copy number for intact elements is neither a necessary nor a sufficient condition for non-autonomous elements to invade. Nevertheless, invasion occurs under a broad range of models and conditions. This requires neither that the non-autonomous elements produce a trans-acting repressor of transposition, nor that they titrate transposase. The elimination of autonomous elements follows the increase in non-autonomous elements unless the latter encode powerful repressors of transposition. Approximate solutions for the equilibrium copy number of autonomous elements and rate of invasion of non-autonomous elements can be found under some models for transposition and selection. The predictions of the model are compared with recent empirical studies of the D. melanogaster P system

Mitochondrial chloramphenicol and oligomycin resistance mutations were used to investigate mitochondrial inheritance in A. nidulans. Mitochondrial RFLPs could not be used to distinguish between paternal and maternal mitochondria because none were detected in the 54 isolates investigated. Several thousand ascospores from each of 111 hybrid cleistothecia from 21 different crosses between 7 heterokaryon incompatible isolates were tested for biparental inheritance. All mitochondrial inheritance was strictly uniparental. Not one instance of paternal inheritance of mitochondria was observed. The implications of our results for the theory that uniparental inheritance evolved to avoid cytoplasmic conflictare discussed. Possible explanations for the maintenance of strict uniparental inheritance of mitochondria in an inbreeding homothallic organism are suggested. The chloramphenicol resistance marker was inherited preferentially to the oligomycin resistance marker probably due to the inhibited energy production of mitochondria with the oligomycin resistance mutation. The maternal parent was determined for 93 hybrid cleistothecia from 17 crosses between 7 different strains. Contraryto previous reports A. nidulans strains functioned as both maternal and paternal parent in most crosses.

Microsatellites closely associated with each member of the Tcp1O gene family were amplified simultaneously from t haplotype and wild-type forms of mouse chromosome 17, by PCR. The t complex responder (Tcr) locus, which plays a central role in transmission ratio distortion, maps within the Tcp10 cluster on the t haplotype. Thus the amplified set of microsatellite loci (referred to collectively as Tcp10ms) provides a direct marker for this central component of the meiotic drive system associated with all naturally occurring t haplotypes. A unique Tcp10ms pattern of microsatellite alleles was obtained for a number of independent, laboratory-maintained complete and partial t haplotypes. Independent t chromosomes found in wild mice from US populations also had unique patterns, even when they were classified within the same lethal complementation group. Wild and laboratory chromosomes in the tw5 group showed similarly-sized but non-identical Tcp10ms patterns, suggesting they share a recent common ancestor. These chromosomes are likely to have derived from an ancestral chromosome within the founding population of North American house mice. The Tcp10ms pattern was also shown to be useful in field studies for distinguishing among independentt haplotypes, when more than one is present within a single population.

In this report we mapped a new MEL11 gene and summarize our population studies of the α-galactosidase MEL genes of S. cerevisiae. The unique family of structural MEL genes has undergone rapid translocations to the telomeres of most chromosomes in some specific Saccharomyces cerevisiae populations inhabiting olive oil processing waste (alpechin) and animal intestines. A comparative study of MEL genes in wine, pathogenic and alpechin populations of S.cerevisiae was conducted using genetic hybridization analysis, molecular karyotypingand Southern hybridization with the MEL1 probe. Five polymeric genes for the fermentation of melibiose, MEL3, MEL4, MEL6, MEL7, MEL11, were identified in an alpechin strain CBS 3081. The new MEL11 gene was mapped by tetrad analysis to the left telomeric region of chromosome I. In contrast, in wine and pathogenic populations of S. cerevisiae, MEL genes have been apparently eliminated. Their rare Mel+ strains carry only one of the MEL1, MEL2, or MEL8 genes. One clinical strain YJM273 was heterozygotic on the MEL1 gene; its mell0 allele did not have a sequence of the gene.

7-tricosene (7-T) and 7-pentacosene (7-P) are the two main hydrocarbons on the cuticle of male Drosophila melanogaster. These two substances might play a pheromonal role during courtship behaviour. We investigated the genetic basis of the quantitative polymorphism observed in the production of 7-T and 7-P. Strains of different geographic origin, with males producing either predominantly 7-T or predominantly 7-P, were hybridized with strains carrying genetic markers. We found that chromosome II changes the balance between 7-T and 7-P while chromosome III regulates the overall quantity of both 7-monoenes. We have also characterized and roughly mapped sept and smoq, two genetic factors on chromosome II that act additively on the production of both cuticular hydrocarbons. The genetic control of the variation in 7-T and 7-P varies between D. melanogaster strains and between D. melanogaster and its sibling species D.simulans. The possible evolutionary and physiological causes of this variation as well as its functional implication for courtship behaviour are discussed.

Currently in the hobo system of hybrid dysgenesis, strain classification is based on the presence/absence of the 2·6 kb Xho I restriction fragment. Using this criterion, strains are classified as: (1) H strains when full-size elements are detected by presence of a 2·6 kb Xho I restriction fragment; they can also contain internally deleted elements; (2) DH strains when only deleted elements are detected (Xho I restriction fragment less than 2·6 kb); (3) E strains, devoid of any restriction fragment equal to or less than 2·6 kb in length. In addition, the strains can be classified on their ability to generate gonadal atrophy (GD sterility) when males of a studied strain are crossed with females from an E strain (dysgenic cross). Here we try to define the nature of the dysgenic cross, which leads us to analyse the different components of the dysgenic syndrome and to look for eventual correlations between them. Molecular analysis, GD sterility tests, hobo mobilization with the haw strain and the vgal strain, and hereditary transmission of the instability at the vg locus have been assayed in different strains. We show that the occurrence of GD sterility depends on the tested H strains as expected, but also on the E strains used. On the other hand we do not find any correlation between the different dysgenic parameters. Our data reveal that molecular and GD sterility tests are not sufficient to classify strains in the hobo system, and that all the components of the dysgenic syndrome must be taken into account. Our results are discussed with regard to active and full-size elements in relation to the structure of the S region where an amino acid sequence (TPE) presents a repetition polymorphism

Type I repressors control P element transposition and comprise full length elements and elements with small 3′ deletions in the final exon. Using a sensitive assay for measuring the strength of repression of P element transposition in somatic and germline tissues, we have isolated and characterized a naturally occurring type I repressor element from a Q population of Drosophila melanogaster. We demonstrate that the almost complete repression of transposition in this population is a mixture of KP elements with intermediate levels of repression, and the strong contribution of a single 2·6 kb P element deletion derivative, which we call SR (Strong Repressor). A deletion in the final intron of SR allows for the constitutive production of a putative 75 kDa repressor protein in germline tissues in addition to the production of the 66 kDa repressor in the soma, which would result in a biparental mode of inheritance of repression. Basedon the four observed classes of natural Q populations, we propose a model in which populations containing SR-like elements, capable of producing strong type I repressor constitutively, have a selective advantage over populations which rely either on maternally transmitted P cytotype or on KPinduced weak levels of repression. Such populations may subsequently spread and constitute an evolutionary stable strategy for the repression of hybrid dysgenesis in Drosophila melanogaster.

The species cross between Drosophila melanogaster and D. simulans was first described by Sturtevant in the 1920s. According to his description, the hybridization of D. simulans females and D. melanogaster males produces only (or almost only) male progeny. Female hybrids are embryonic lethal. Here it is shown that these traditional results no longer hold. Instead, D. simulans is polymorphic for factor(s) that qualitatively affect the outcome of species crosses to D. melanogaster. Remarkably, many, if not most, strains of D. simulans produce abundant female hybrids when crossed to D. melanogaster males. Genetic analysis of the difference between D. simulans strains that produce many versus few hybrid females shows that recovery of hybrid females depends on autosomal, maternally acting gene(s).

In this paper we report on the diversity and distribution of a set of non-LTR retrotransposon (RTP) reverse transcriptase (RT) sequences isolated from phlebotomine sandflies, and their potential for investigating the evolutionary histories of members of this subfamily of flies (Diptera: Psychodidae, Phlebotominae). The phlebotomine RT sequence families derived from one species were as different from each other as they were from RT sequences derived from other species. When each was used to probe Southern blots of sandfly genomic DNA they hybridized only to the species of source and, usually, to others of the same subgenus, but not to DNA from other subgenera — a hybridization pattern consistent with vertical evolution. There was considerable intraspecific variation in hybridization pattern, suggesting the RTs were part of non-LTR RTPs that are (or were recently) subject to flux in genomic position and copy number. Most of the RT families detected in phlebotomines are monophyletic with respect to previously described RTs, and all are monophyletic with RTs of the F/Jockey (Drosophila melanogaster) type of RTP. Orthologous sequences were isolated from the closely related species Phlebotomus perniciosus and P. tobbi (subgenus Larroussius), and different populations of P. perniciosus. The level of sequence divergence among these orthologous RTs, the subgeneric distribution of each RT family, and the intraspecific variation in hybridization pattern of many of them, indicate this class of sequence will provide genetic markers at the sub-generic level.

Gene amplification involving a particular haplotype has been found at the esterase B locus of mosquitoes from various countries. This similarity has been explained by a unique amplification event followed by migration and selection by organophosphate (OP) insecticides. This assumes that the polymorphism of non-amplified esterase haplotypes is so large that the chance of independent amplification in two distinct populations is negligible. In order to test this assumption, three susceptible populations from northern Europe were sampled and analysed for esterase and haplotype polymorphism. At the protein level, 18 and 16 alleles were found for esterase A and B respectively in one French population (n = 74), and 16 and 14 in an English one(n = 50). At the DNA level, 24 alleles at the esterase B locus were detected in a sample of 72 mosquitoes from one population, with the use of only one restrictionenzyme (EcoR V). Restriction maps of two nonamplified haplotypes randomly sampled from a single breeding site in Belgium were built with six restriction enzymes. 60% of all restriction sites were different among the two maps. The huge polymorphism found in northern Europe requires specific explanations for its stability, but it considerably strengthens the hypothesis of migration of amplified haplotypes.

The pattern of chromosomal variation is investigated in house mice from the Danish hybrid zone between the translocation-prone Mus musculus domesticus and the chromosomally conservative M. m. musculus. The cytogenetic analysis confirmed the non-introgression of three pairs of Robertsonian (Rb) fusions from M. m. domesticus into the M. m. musculus genome. The geographic distribution of two of these Rb fusions was shown to follow staggered chromosomal clines which increased in steepness the closer they were to the centre of the hybrid zone as defined by allozymes. Analysis of alternate hypotheses suggests that chromosomal differentiation of the Danish domesticus occurred after contact was established with musculus. The staggering of the clines would reflect the order of arrival of theRb fusions into the hybrid zone. Several models with different processes of underdominance of the chromosomal heterozygotes are discussed to account for the difference in width between clines. A selective model with increasing levels of genomic underdominance due to interaction with a progressively enriched musculus genome provides the best fit for the observed pattern. Selection against Rb fusions with little effect on the recombination of linked allozyme markers supportsthe view that no reduction in gene flow due to chromosomal heterozygosity is yet apparent through the hybrid zone and that only the centromeric segments of the Rb fusions are incompatible with the musculus genome.

Frequencies of anaphase I nondisjunction, germ cell death and pairing abnormalities at pachytene were assessed in male mice singly heterozygous and homozygous for the Robertsonian (Rb) translocations: Rb (1.3)lBnr, Rb(ll. 13)4Bnr and Rb(10. ll)8Bnr. Rb homozygotes showed low frequencies of nondisjunction but substantial germ cell death. This germ cell death could not be attributed to problems at pachytene as Rb homozygotes showed no increase in pairing abnormalities over the (C3H/HeH×1O1/H)F1 controls. Instead genie factors are involved. Rb heterozygotes showed substantial frequencies of nondisjunction and even greater germ cell death than found in the homozygotes. Pachytene pairing abnormalities were observed and it appears that these, together with genie factors, cause physiological perturbation of meiocytes, thereby promoting germ cell death, with nondisjunction of the trivalent as a sublethal response.

The mouse transgene, provisionally designated TgN(Hbb-b1)83Clo, was produced by Dr C. Lo by pronuclear injection of the cloned ß-major globin gene and comprises a highly reiterated sequence that is readily detected by DNA in situ hybridization on histological sections. This fulfils many of the requirements of an ideal genetic cell marker and has been widely used for lineage studies with mouse chimaeras. However, it is not known whether it causes cell selection or influences developmental processes, such as cell mixing, in chimaeric tissues. In the present study, nontransgenic genetic markers (electrophoretic polymorphisms of glucose phosphate isomerase and differences in eye pigmentation) revealed no significant effect of the presence of hemizygous transgenic cells on the overall composition, size or gross morphology of 12½ d chimaeric foetuses, placentas or extraembryonic membranes. Also, a previously described maternal genetic effect on the composition of chimaeric tissues occurred in the presence or absence of the transgene. These tests have demonstrated that hemizygous cells are not at a significant selective disadvantage, when incorporated into mouse aggregationchimaeras with non-transgenic cells. Further studies are needed to test whether homozygous transgenic cells are also selectively neutral and to test whether hemizygous or homozygous transgenic cells influence developmental processes, such as cell mixing, that were not tested.

Experiments on Drosophila suggest that genetic recombination may result in lowered fitness of progeny (a ‘recombination load’). This has been interpreted as evidence either for a direct effect of recombination on fitness, or for the maintenance of linkage disequilibria by epistatic selection. Here we show that such a recombination load is to be expected even if selection favours increased genetic recombination. This is because of the fact that, although a modifier may suffer an immediate loss of fitness if it increases recombination, it eventually becomes associated with a higher additive genetic variance in fitness, which allows a faster response to directionselection. This argument applies to mutation-selection balance with synergistic epistasis, directional selection on quantitative traits, and ectopic exchange among transposable elements. Further experiments are needed to determine whether the selection against recombination due to trie immediate load is outweighed by the increased additive variance in fitness produced by recombination.

In plants, models for mapping quantitative trait loci (QTL) based on flanking markers have been mainly developed for progenies of inbred lines. We propose twoflanking marker models for QTL mapping in F1 progenies of non-inbred parents. The first is based on the segregation of four different scorable alleles at a marker locus (the four-allele model) and the second (the commonallele model) on one scorable allele per marker locus segregating in both parents. These models are suitable for the majority of the allelic configurations which may occur in crosses between heterozygous parents. For both cases, when four scorable or one common-allele per marker locus segregate, additional algorithms were developed to estimate the recombination frequency between two marker loci. Tests carried out with simulated populations of various sizes indicate that the models provide a good estimate of QTL genotypic means and of recombination frequencies between flanking markers and between the marker loci and the QTL.The estimates of QTL genotypic means have a higher precision than the estimates of recombination frequencies. The four-allele model shows a higher ability to detect QTLs than the common-allele model. If segregation ratios are distorted, the power of both models and the precision of the estimates of recombination frequencies are reduced, whereas the accuracy of estimates of QTL genotype means is not affected by distorted segregation ratios. The power of the common-allele model is substantially reduced if QTL genotypic means depend on additive allelic interactions, whereas the four-allele model is less affected by the non-additive behaviour of QTL alleles.

The t-complex on mouse chromosome 17 results in transmission ratio distortion in males heterozygous for complete haplotypes, and sterility in those homozygous for semi-lethal or doubly heterozygous for complementing lethal haplotypes. This sterility is due to inability of spermatozoa to fertilize. The haplotype th7 is an unusual laboratory-derived haplotype, postulated to carry a small duplication of t chromatin. Males heterozygous for th7 show a new form of sterility, apparently due to failure to form copulation plugs during mating. This is accompanied by a strong propensity to acute urinary obstruction. It is suggested that both the failure to form copulation plugs and the urinary obstruction are due to some abnormality in function of the accessory sex glands, and are the result of incorrect dosage of a gene in the postulated duplication. The symbol Msu for male sterility and urinary obstruction is suggested for the locus concerned. Previously a recessive form of abnormal behaviour had also been attributed to this duplication.

A new method is presented for estimating the rate of gene flow into island populations using the distribution of alleles in samples from a number of islands. The pseudo maximum likelihood estimator (PMLE) that we derive may be applied to species with either discrete or continuous generation times. For Wright's discrete-generation island model, the method provides an estimate of θ = 2Nm where N is the (haploid) population size on each island and m is the fraction of individuals replaced by immigrants in each generation. For a continuous-generation island model, the corresponding parameter φ is the ratio of the immigration rate φ to the individual birth rate λ. Monte Carlo simulations are used to compare the statistical properties of the PMLE with those of two alternative estimatorsof θ derived from Wright's F-statistics. The PMLE is shown to have greatest efficiency (least mean square error) in most cases for a wide range of sample sizes and parameter values. The PMLE is applied to estimate θ using mtDNA haplotypes and allozymes for subdivided populations of African elephants and Channel Island foxes.

The recombination frequency (r) between two loci defined by conventional or molecular markers can be estimated by solving proper Maximum Likelihood equations. These are based on expected and observed marker class frequencies in the progeny of a cross, and are specific for each allelic configuration of the parents(1). In a cross between two diploid parents up to four different alleles, besides a null allele, can be detected at one locus. This defines in each parent, considering a locus A, nine basic allelic configurations based on two allelic marker fragments(Ai/Aj), one single marker allele and a null allele (Ai/AO), or just null alleles (AO/AO). With respect to two loci the consideration of all possible diploid allelic configurations in the parents of a cross allows the detection of 21 different expected marker class distributions producing estimates of r in the progeny. General formulas for calculating the ML equations and the corresponding information functions have been developed for the 21 marker class distributions. Simplified formulas have been also derived and the relative efficiency of the information functions compared. As expected, in the majority of cases, allelic marker configurations give more precise estimates of linkage values than single marker configurations. A method for the construction of linkage maps based on two point estimates, linkage subgroups and allelic bridges is presented. The method is an improvementon an original proposal by Ritter et al.(1990).

We investigated the interaction of recombination and selection on the process of fixation of two linked loci with epistatic interactions in fitness. We consider both the probability of fixation of newly arising mutants (the static model) and the time to fixation under continued mutation (the dynamic model). Our results show that the fixation of a new advantageous combination is facilitated by higher fitness of the advantageous genotype and by weaker selection against the intermediate deleterious genotypes. Fixation occurs more rapidly when the recombination rates are small, except when selection against intermediate genotypes is weak and selection in favour of the double mutant is very strong. In these cases fixation is more rapid when the recombinant rate is large. Mutations of strong effects, deleterious when alone but beneficial when coupled, are fixed more easily than mutations of intermediate effects, at least for large recombination rates. Among the possible pathways the process of fixation might follow, independent substitutions lead to the fixation of the double mutant only when selection is weak. The relative importance of the other pathways depends on the interaction between recombination and selection. The coupled-gamete pathway (i.e. when the population waits until the double mutant appears and then drives it to fixation) is more important as selection intensity increases and the recombination rate is reduced. For all recombination rates, asymmetries in fitness of the intermediate genotypes increase the rate at which fixations occur. Finally, throughout the fixation process, the population will be monomorphic at least at one of the two loci for most of the time, which implies that there would be little opportunity to detect the presence of negative epistasis even if it were important for occasional evolutionary transitions.