ALLIED TO LACERTA MURALIS. 11 



7. The rather large, rhombic or hexagonal, keeled dorsal scales of L. agilis lead 

 through various gradations to the smooth grannies of L. ocellata, and of many of the 

 members of the L. muralis group ; a tendency to imbrication, more or less marked in 

 L. agilis and L. viridis, leads to L. princeps and, with an increase in size, to Algiroides 

 and Psammodromtfs. Tlic distinctly overlapping character of the scales and shields of 

 the lower parts in L. agilis is lost in the Z. mnralis group, and the number, 6 or 8, 

 of longitudinal rows of these shields, gradually merging into the lateral scales, may 

 be increased, as in L. ocellata and galloti, or the differentiation may be more abrupt 

 through disintegration of the small outer shields as in many members of the L. muralis 

 group. The large plates forming the so-called collar may be reduced in size and 

 increased in number, at the same time losing the notchps between thera on their free 

 border, as in L. muralis and allies, or they may become adherent to tlie middle of the 

 breast, as in Acanthodacti/lus ; or the whole collar may lose its freedom, and the gular 

 scales pass gradually into the ventral ])lates, as in Fsammodromus. 



S. Of all the species of Lacerta, L. agilis has the shortest digits, and they are not 

 at all compressed. In the L. muralis group the digits become longer and more 

 compressed, especially distally, as the climbing habits become more and more marked. 

 The series L. taurica-campestris-serpa is instructive in this respect. In L. agilis the 

 subdigital lamelUe are smooth and undivided, or divided into two. In Lizards adapted 

 for arid sandy tracts the subdigital lamellae acquire one or several keels {Latastia, 

 Acanthodacty I us, Eremias), and a series of pointed scales may form a serration or 

 fringe on one or both sides of the digit {Acanthodactglus). These are clearly adapta- 

 tions to a special mode of life. I cannot conceive the direction of the series reversed, 

 viz., compressed or serrated digits leading to the simple form of X. agilis. 



9. If ontogeny is a guide to phylogeny, L. agilis, which has the shortest tail in the 

 genus Lacerta, must be regarded as tlie most primitive species in this respect, Lizards 

 at birth having invariably a shorter tail in proportion to the body *. Mehely, who 

 considers the longer tail as the more primitive, observes that extremely long, slender, 

 whip-like tails — as in L. od-yeephala, for instance — are endowed with greater fragility. 

 This may be true — and here again I would look upon extreme fragility of the organ 

 as a specialization, — but I must say that in L. agilis, which has a short and thick tail, 

 individuals with the organ in a regenerated condition appear to me as frequent as in 

 most other species of the genus, and more so than in L. echinata, whicli has the 

 longest and most whip-like tail (spinose at the base, hence the name Centromastix). 



lU. According to Eimerf, whose views have been endorsed by Cope;}: and by 



• Measurements of head and body and tail in specimens at birth of four species : — L. agilis, 26H-i'9 mm. : 

 L. viridis, '2ij-\-27 ; L. vivi^iara, 20 + l'2 ; L. muralis, 22 + 28. 

 t Arch. f. Naturg. xlvii. i. p. 239, pis. xiii.-xv. (18^1). 

 J llcp. U.S. Nat. Mus. f. 1S98, p. 509, pi. v. (1900). 



c2 



