FLOKAL YARTATION IN YERONICA PERSICA 353 



Six (lowers witli a diiiierons corolla were found. The calj^x was 

 normal in each case, cv)nsi.sting' o£ two postei'ior and two larger 

 anterior segments. Both corolla-lobes were subtruncate at the base, 

 and broader than long, the posterior lobe being dark blue, and the 

 anterior one paler and slightly lai-ger. Measurements from two 

 llowers were as follows : posterior lobe 3*5 mm. long, 5 5-6' 5 mm. 

 broad ; anterior lobe 3*5 mm. long, 7 mm. broad. The posterior lobe 

 aj)[)arcntl3^ represented two petals and the anterior one three petals only, 

 as is the case in many bilabiate GiiDiopetalce. Jules Camus stated — 

 and I made the same observation in Somerset before reading his paper 

 — that the posterior corolla-lobe in a normal flower of V. persica has 

 no median nerve, there being two nerves near the middle, equidistant 

 from the median line. The lateral and anterior lobes, on the other 

 hand, have a median nerve. The absence of a middle nerve in the 

 posterior corolla-lobe is, however, not constant. In Hertfordshire a 

 middle nerve seems to occur just as frequently as not. But the fact 

 that it is sometimes absent \n-Aj perhaps be regarded as affording an 

 indication of the double nature of the posterior lobe. That both 

 lobes of the dimerous corollas represented more than one petal each 

 was suggested by their breadth and the relatively large number of 

 nerves — 10-11 in the posterior lobe and 12-13 in the anterior one. 

 The posterior lobe had no middle nerve, and may be regarded as 

 composed of two petals ; the anterior lobe, being slightlj^ larger and 

 possessing a middle nerve, probably represented three petals. The 

 nerves of the posterior lobe were not forked, whereas 4-5 nerves of 

 the anterior one were consj^icuously forked. The assumption seems 

 justified that the " dimerous " corollas observed by me were composed 

 ot" live petals, one more than are present in the usual trimerous type. 

 Penzig considered that only four petals were represented in the 

 "dimerous" corollas examined by him, but gave no reasons for re- 

 garding the anterior lobe as being composed of two petals, rather than 

 of three (Ptlanzen-Terat. ed. 2, iii. 121). It should be mentioned 

 that one half of a corolla-lobe in V. i^ersica sometimes possesses one 

 nerve more than the other half; hence there is sometimes an odd 

 number of nerves in a lobe without a median nerve and an even 

 number in a lobe possessing a median nerve. 



The dimerous corollas and a great majority of the 105 trimerous 

 ones were on starved plants, a few inches high, growing at the edge 

 of the mangold- field, whei-e there was a rough path and the soil was 

 consequently trodden dovvn and stiff. All the flowers in which addi- 

 tional sepals, corolla-lobes, or anterior staminodes occurred were borne 

 by reJativelij vigorous plants in the rest of the mangold- field and the 

 kitchen-garden, where the soil was comparatively loose. Some of 

 these plants had about six stems, each about 18 inches long. 



Bateson, wlio examined 1328 flowers of V. i)ersica from plants 

 grov/ing in stubbles on heavy land round Cambridge found 93 (7 per 

 cent.) with trnnerous corollas and 14 (1 jjer cent.) with dimerous 

 ones, only one corolla having an additional petal. In a plot of waste 

 garden-land, on the other hand, 10 flowers out of 286 (3'5 per cent.) 

 had five corolla-lobes, the remainder being normal ( Journ. Linn. Soc, 

 Bot. xxviii. 397). Taken in conjunction with my observations, these 



