Research Article |
Corresponding author: Tarombera Mwabvu ( tarombera.mwabvu@ump.ac.za ) Academic editor: Michelle Hamer
© 2022 Tarombera Mwabvu, Adekunle Adebowale.
This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Citation:
Mwabvu T, Adebowale A (2022) Geometric morphometric analysis of gonopods in Bicoxidens flavicollis populations (Diplopoda, Spirostreptida, Spirostreptidae). African Invertebrates 63(1): 77-88. https://doi.org/10.3897/afrinvertebr.63.81092
|
Male gonopods are useful in taxonomic diagnoses and descriptions of millipedes, although they may vary intraspecifically in shape and size. To assess this intraspecific variation, we used geometric morphometric analysis to compare gonopod morphology among eight isolated populations of the colour-polymorphic southern African millipede Bicoxidens flavicollis. Our results showed that gonopod shapes vary significantly across the examined populations, and elucidated subtle variations. CVA cross-validation test indicates an average classification rate of 75% for the five populations for which we had more than one specimen. Although we had a small number of replicates for three populations, our results still illustrate the importance of applying quantitative approaches to what would otherwise be qualitative and subjective gonopod shape categories in millipedes. As such, the taxonomic assignment of the populations of B. flavicollis may require further investigation, and further revisions would be required with an integrative approach, including molecular data, in order to re-evaluate the taxonomic diversity and distribution data of this species. Finally, we highlight the conservation potential of divergent populations as evolutionary insurance against a dynamic and unpredictable climate, whether or not they undergo full speciation.
cryptic species, millipedes, morphology, taxonomic diversity
Morphological differences are important in taxonomy to delimit taxa (
The genus Bicoxidens Attems, 1928 comprises nine species which are endemic to southern Africa, occurring in diverse habitats such as Miombo woodland, riverine, and montane vegetation (
Given the colour polymorphism and high genetic divergence (18% for CO1 and 6% for 16S rRNA) recorded in two populations of B. flavicollis (
From this perspective, we employed in this study for the first time a geometric morphometric analysis to assess the taxonomic value of gonopod morphology in separating populations, and to test the hypothesis that gonopod morphology may underestimate the taxonomic diversity of a spirostreptidan species.
Copulating and non-copulating males and females of B. flavicollis were collected by hand from eight geographically separate populations in Zimbabwe as summarised in Table
Population (sample size) | Coordinates | Province | Voucher | Vegetation/Habitat |
---|---|---|---|---|
Chegutu (1) | 18°06'S, 30°09'E | Mashonaland West | NM 21954 | Dry savanna |
Chihota (1) | 18°16'S, 31°05'E | Mashonaland East | NM-Myr 25833 | Miombo woodland |
Chitombo (12) | 18°26'S, 32°58'E | Manicaland East | NM-Mil 25832 | Miombo woodland |
Marange (4) | 19°10'S, 32°18'E | Manicaland West | NHMZ | Dry savanna |
Mazowe (5) | 17°29'S, 30°59'E | Mashonaland Central | NM 21958 | Miombo woodland |
Muterere (1) | 18°26'S, 32°58'E | Manicaland East | NM-Mil 25831 | Miombo woodland |
Muzinga (5) | 18°25'S, 32°58'E | Manicaland East | NM-Mil 25835 | Miombo woodland |
Sahumani (11) | 18°32'S, 32°50'E | Manicaland East | NHMZ | Miombo woodland |
All the specimens were preserved in 96% ethanol. The voucher specimens from each population were deposited in the Natural History Museum (NHMZ), Bulawayo, Zimbabwe, and the KwaZulu-Natal Museum (
All the samples were photographed in the same oral view and magnifications. In total, eighteen landmarks for 40 specimens were obtained from the examined gonopods (Table
where R is the repeatability value, is the among-individuals component of variance, and is the within-individuals variance component (Sokal and Rohlf 1995; Arnqvist and Martensson 1998; Fruciano 2016).
For shape analysis, the variance-covariance matrix derived from the Procrustes coordinates were subjected to exploratory principal component analysis to reduce the dimensionality of the data to the most significant shape variables. As the first 15 principal components (PCs) accounted for about 95% of observed variance, the corresponding PC scores were selected as variables for a canonical variates analysis (CVA) to assess shape discrimination across populations. This was followed with cross-validation tests to assess the rate of correct classification of specimens into their population groups. We tested the null hypothesis that gonopod shapes are not different among the eight populations of B. flavicollis using the PC scores and population as dependent and predictor variables, respectively. All analyses were implemented in PAST (Hammer et al. 2001) and MorphoJ (Klingenberg 2011).
Description of B. flavicollis gonopod landmarks and semi-landmarks used for shape analysis.
Landmark number | Landmark definition (following |
Landmark type (after |
---|---|---|
1 | Apex of median lobe of apical proplica | I |
2 | Apex of axe-shaped process (or apical axe-shaped process) | I |
3 | Tip of lobe of axe-shaped process | I |
4 | Base of lobe of axe-shaped process | I |
5 | Lateral mid-proplica | II (semi-landmark; midpoint between landmarks 4 and 6) |
6 | Apex of sternite (or apical sternite) | I |
7 | Base of sternite (or basal sternite) | I |
8 | Base of proplica (or basal proplica) | II (semi-landmark; the most distal part of the basal prolica) |
9 | Basal lateral edge of paracoxite | II (semi-landmark; furthest point on the lateral edge of paracoxite |
10 | Distal lateral paracoxite | I |
11 | Paracoxite apex | I |
12 | Telopodite (at midlength between knee and femoral lobe) | II (semi-landmark; midpoint between knee and femoral lobe) |
13 | Base of rounded lobe on proplica | I |
14 | Base/pit of telopodite knee | I |
15 | Apex of telopodite knee | I |
16 | Proximal edge of lateral process | I |
17 | Apex of lateral process | I |
18 | Apical groove between lateral process and median lobe | I |
The assessment of measurement errors arising from landmark digitisation demonstrates that variation between repeated digitisations of the same specimen is significantly lower than variation among different specimens, with an intraclass correlation coefficient of 0.99. The thin plate spline deformation grids (Fig.
Although a scatter plot of specimens along PC1 and PC2 (Fig.
Specimens from Chitombo, Marange, Mazowe, Muzinga, and Sahumani appear to occupy distinct morpho-space with minimal overlap. Despite the low sample size, cross-validation tests showed a relatively high percentage of correct specimen classification for populations with more than one specimen as follows: Chitombo (83%); Marange (75%); Mazowe (60%); Muzinga (75%); Sahumani (82%), and an overall classification rate of 75%.
For any morphological structure in an organism, there are almost limitless possibilities of theoretical shapes that could evolve. However, only a few are evolutionarily viable due to functional, and thus adaptive constraints placed on such structures by natural selection (
Given this background and the significance of male gonopods in the reproductive success of arthropods, some levels of morphological stasis in gonopods would be expected in a species as widely distributed as B. flavicollis. Our results demonstrate that gonopod shape in B. flavicollis varies significantly among the eight populations, with landmarks 6, 7 and 11 showing pronounced shape changes along the first two PCs. However, it is conceivable that extensive sampling of Chihota and Muterere specimens would have blurred the distinction between them and their respective nearest neighbour population groups. The combination of landmarks 6 and 7, corresponding to the apex and base of the sternite, defines a spatial relation between two points that could be discriminatory at interspecific level of comparison. These three landmarks (6, 7, and 11) are regarded as type I (see
Based on the similarities of gonopods in a widely distributed African millipede genus Doratogonus Attems 1914,
Although observations of conservatism in morphology have been reported in a number of animal taxa including spiders (
Future work on B. flavicollis should include the use of nuclear markers to compare the levels of genetic divergence among the populations. Based on the similar pattern of body colour between populations, the high levels of genetic divergence reported by
Even though three out of eight populations in our study had one specimen, our study highlights the importance of using quantitative methods in taxonomy. Our results support the position that population divergence and variation in male genitalia of B. flavicollis could be coupled.
We wish to thank the KwaZulu-Natal Museum for granting access to the millipede collection, University of KwaZulu-Natal (Microscopy Unit, Pietermaritzburg) for making photographic equipment available and the National Research Foundation—through the Knowledge, Interchange and Collaboration (KIC) (Grant No. 92969)—for providing research funds to T. Mwabvu. We are grateful to Takudzwa Samuneta and Denounce Mukora for assisting in fieldwork.