Estimates of gene flow between tsetse populations based on population genetic analyses are important in the context of predicting the long-term effectiveness of tsetse control measures . However, genetic markers also have other uses, in particular as tools that can assist in determining the genetic basis underpinning important phenotypes. When captive colonies are used to analyse phenotpyes such as factors influencing vector competence [22, 23] it is important that this can be related back to the field populations.
This study demonstrates that AFLPs can be used to generate high-resolution genetic markers in natural populations of G. m. morsitans. Two distinct population groups were apparent from the cluster and PCA analysis (Figures 2 and 3). Although the separation of the two populations is not perfect, the results demonstrate that the AFLP markers can be used to separate two genetically close populations. The level of Nei's genetic diversity , h was small in both populations and the amount of genetic variation detected in the laboratory and a Zimbabwean population was similar. In addition, the GST value of 0.0386 indicates that there is little genetic differentiation between the natural Zimbabwean and the laboratory population. The results in the current study suggest that the establishment and long term maintenance of the insect colony process (> 30 years) has resulted in little effect on genetic variation as estimated by AFLPs. The fact that the flies from the laboratory colony exhibit only limited divergence from the field populations is encouraging for extrapolation of experimental results to the 'real world'. The field population did, however, show a higher number of loci and a higher percentage of polymorphic loci than the laboratory population (Table 1). AMOVA results showed that the populations are closely related as there was more variation within (91%) the populations than between (9%) them. This result, derived from variation in nuclear genes, is in agreement with that obtained for G. morsitan morsitans using mitochondrial markers . It was interesting that males and females were clearly separated using the AFLP analysis, however this can be explained by the greater DNA content of female flies where the dosage of X determines sex. Hence males are XY or XYY whereas females can be XX, XXY or XXXY . The identification of markers that can discriminate males and female flies could be valuable in the future in the identification of sex ratio distortion genes  in the tsetse genome.
Whenever possible it is preferable for DNA to be extracted from fresh tsetse material. However, when this is impractical due to location, cost or a combination of the two, the present results suggest that acetone should be used in preference to the other preservatives tested. Flies preserved in acetone gave the highest quality HMW DNA and resulted in DNA of good quality on both fresh and pre-dried material. Ethanol has been widely used for the preservation of various biological samples including insects [26–29]. In the current work both tsetse DNA yield and purity from ethanol samples were similar to acetone preserved material (Figure 1). However, ethanol preserved samples contained a low amount of HMW DNA and ethanol was inferior in the preservation of fresh flies (Figure 1). This observation is consistent with previous studies that demonstrated poor ethanol preservation in the presence of water [26, 28]. This problem was not overcome by using 100% proof ethanol and changing the ethanol after 24 h (this study).
Although the material preserved in SDS lysis buffers (either with or without proteinase K) or on a FTA card showed reasonable DNA yield and purity in all cases it was degraded LMW DNA with pronounced smearing that predominated. The LMW smearing was most extensive in the FTA samples suggesting that FTA cards are not appropriate media for storing large insects. This may be because not all parts of the fly were in contact with the matrix, allowing endonucleases activity resulting in DNA degradation.
The data presented here shows that acetone was a superior medium for preserving tsetse DNA. An additional advantage is that acetone is used as a bait component in tsetse traps and is therefore readily available to field workers collecting tsetse flies. The DNA obtained using this method could be used for PCR of specific genes, microsatellite analysis, AFLP and other molecular techniques that require high quality DNA. In addition, acetone may be an appropriate storage media for other important dipteran vectors, such as mosquitoes due to its high penetrability and dehydrating abilities. However, one of the disadvantages of using acetone is that many aircraft companies will not allow acetone to be carried in the hold.
The relative value of a range of genetic markers, specifically Random Amplification of Polymorphic DNA (RAPDs), Inter-simple-sequence repeat PCR and microsatellites as applied to the analysis of variation between strains of the silkworm (Bombyx mori) has been reviewed , although this study did not include AFLPs. The basic conclusion was that all techniques are useful and have different strengths and weaknesses in terms of factors such as resolution, ease of use and cost. As regards nuclear genetic markers in tsetse flies, allozymes sample approximately 50 loci and do reveal a degree of inter-population diversity, however they are not selectively neutral and hence may not always accurately reflect population history. Microsatellites reveal significant allelic polymorphism but are relatively few in number and are limited by the requirement for sequence data for primer design. Both have proved useful for initial analysis of inter-population differentiation in the field [11, 12]. The initial AFLP study described here has identified numerous previously unknown polymorphic loci in G. m. morsitans which as indicated above will have value in areas of tsetse research, additional to differentiation of field populations. Because these loci are essentially randomly distributed , greater in number and not necessarily selected for rapid evolution, AFLP analysis may provide a more accurate assessment of genome-wide differentiation in Glossina. In future it should prove possible to generate microsatellite markers more rapidly as expressed sequence tag and genome sequence databases  are systematically searched for variable number tandem repeat (VNTR) and single nucleotide polymorphisms (SNP). However, due to the large size of Glossina genomes  it seems likely that not all of these will have genome sequences available in the immediate future. AFLP markers which can generated over a wide dynamic range of genome sizes  should continue to be of value.