TZ 
can be used to generate hypotheses about the habitats and locomo- 
tory modes of species where these are uninvestigated or incompletely 
so. Thus the one known specimen of Mesalina ercolinii occurs in the 
area of Fig. 3 mainly occupied by ground dwelling forms using 
dense vegetation, an exceptional habitat for an advanced armatured 
lacertid. The African Poromera fordi has many morphological re- 
semblances to the east Asian grass lizards, Takydromus, that seem to 
be related to climbing in vegetation (Arnold, 1997) but, although the 
limb pairs of Poromera are not very disparate in length, as expected 
in a climber, they are distinctly longer than in Zakydromus and other 
scansorial species. This suggests a locomotory difference between 
the two genera and perhaps indicates that, although Poromera does 
climb in vegetation, it is also frequently active in open situations, for 
instance it may run on the ground more extensively than Zakydromus. 
Sexual dimorphism in limb length 
It will be seen from Table 1 that there is sexual variation in relative 
length of the hindlimbs, which nearly always appear to be shorter in 
females. In most cases the apparent difference is slight, but in a 
number of taxa, it is substantial, the mean adult male hindlimb span 
in terms of body length sometimes being as much as 12% more than 
that of females. Marked sexual difference occurs in, among others, 
Psammodromus hispanicus, Lacerta agilis, L. lepida, L. andreansz- 
kyi, L. laevis, L. danfordi, Algyroides nigropunctatus, Lacerta 
perspicillata, Adolfus alleni, Podarcis, Philochortus intermedius, 
Latastia longicaudata, Pseuderemias mucronata, Mesalina and 
Ophisops. In many cases, reduced hind limb span in females is 
associated with raised forelimb/hindlimb ratio, so sexual differences 
0.6 
0.5 
0.8 1.0 1.2 
E.N. ARNOLD 
within species follow the general trend among species (Fig. 4). 
There appear however to be exceptions to this regularity, for instance 
in Lacerta vivipara. 
The sporadic distribution of marked sexual difference in limb 
length indicates that it has arisen a number of times. There are also 
phylogenetic indications that sexual dimorphism may have often 
developed by change in limb proportions of the females rather than 
the males. 
The clear relationship among species between limb proportions 
and the kind of spatial niche occupied suggests that intraspecific 
sexual differences in limb length may reflect differences between 
the sexes in microhabitat, although these do not seem to have been 
systematically looked for. In some cases limb dimorphism is often 
associated with differences in dorsal colouring and pattern that may 
possibly be related to the problems of camouflage in different 
environmental situations. Thus, in Podarcis, females not only have 
shorter hind legs but are more obviously longitudinally striped than 
males, a pattern that may be more cryptic in more vegetated situa- 
tions. 
Many forms with sexual dimorphism in limb length also show 
dimorphism in head size which is probably associated with male 
combat for territory and females, large heads presumably conferring 
advantage in this situation. It might be thought that large limbs 
would also be beneficial in this context, but the relationship between 
head and limb size is not precise and some forms where the males 
have large heads show little apparent limb difference between the 
sexes, for instance Lacerta viridis and L. trilineata. The fact that 
sexual dimorphism in limb proportions may be produced by devia 

1.4 1.6 
Fig. 4 Sexual differences in limb proportions in selected dimorphic species. Vertical axis: FL/HL — Forelimb span/hindlimb span. Horizontal axis: HL/SV 
— Hindlimb span/snout—vent distance. Lines join means for the two sexes, females are always to the left. Letters refer to species as indicated in the 
caption of Fig. 3. Sexual differences in hindlimb length are often large, compared with mean species differences; females often have more equal limb 
pairs than males. 
