V(irl(ttion of the Edible Frofj. 251 



oxteiKling to the hip (fig. 5, a), others have si detached posterioi- 

 part as in the typical form (c), \vhil.'=t others again (b) con- 

 nect the two conditions, the posterior part of the fold, though 

 deviating, being confluent with the anterior and forming a 

 bend before reaching the thigh. 



Bolkav nieiitioii-! among the specific differences between 

 Ji. esculent'i, R. lid'bnndd, and I{. chinensh, that tiie dorso- 

 lateral fold is wider (as wide as the upper eyelid) in the 

 second than in the two others. This character is absolutely 

 worthless, for in specimens of the typical form from France 

 and Switzerland its width usually measures ^ to § that of 

 the upper eyelid, but may be equal to it (St. Malo, Havre, 

 liasle, Zofingen), and in German and iVnstro- Hungarian 

 specimens of the var. ridibunda f to | that width is by no 

 means unfrequent. Tlie told is always narrower than the 

 upper eyelid in the vars. chinensis and lesaonce. 



In my previous descriptions of the var. ridibunda I have 

 drawn attention to the fact that the dorso-lateral fold, though 

 usually broader than in the other forms, is less prominent ; 

 I should add that it is sometimes so flat that it cannot be 

 traced without the use of a lens, when the j)ores with which 

 it is studded indicate its course. It has not been pointed out 

 however that these folds are rendered more inconspicuous 

 still owing to the spots on the body being disposed quite 

 inespective of them, whilst in the typical form and the vais. 

 lessonce and cliinensis they stand out on account of their 

 lighter colour, hardly ever encroached upon by the spots, 

 Avliich may he arranged more or less in relation to them, 

 especially when formino- longitudinal bands. When a 

 specimen of the var. ridibunda is seen at a short distance 

 there is usually nothing to reveal the presence of the dorso- 

 lateral folds, wiiich strike the eye in the typical form and 

 the vars. les>ionce and chinensis. 



These facts have a bearing on the question of the derivation 

 of the forms which constitute the species li. esculentUy and 

 confirm ihe view I have held ever since I took up the study 

 of the subject that the var. ridibunda is the most primitive 

 form, out of wiiich the others have been evolved. In a paper 

 recently published * on the derivation of characters in the 

 genus liana as a whole, the absence of the dorsu-lateral fold 

 is considered by me as the primitive condition, and the 

 North American 11. catesbiana^ in which it is totally absent, 

 is, for this and other reasons, regarded as nearest the hypo- 

 thetical ])rototy|je among all the species of Eurasia and 



* iJulI. .-soc. Zool. Frai.ce, 1918, p. 111. 



