230 ANGIOSPERMAEDICOTYLEDONES 



protandrous, very rarely protogynous. As a rule, self-fertilization alone takes place 

 in this form. Insect-visits are extremely rare, even when nectar-guides are present. 



2. Flowers usually bilaterally symmetrical (zygomorphous) to a marked degree, 

 and very large, their diameter being 12-15 mm. Nectar-guides may be present, 

 and if so are clearly marked, moderately large, often nearly bisected by the mid-ribs, 

 and covered with numerous dark streaks and dots. The flowers are markedly 

 protandrous, and self-pollination is usually excluded. 



Schulz says that, at Halle, the flowers of pimpinellifolium are almost always 

 larger than the average of those borne by genuinum, bilaterally symmetrical as a rule, 

 and possessed of nectar-guides, though more rarely they are radially symmetrical, 

 in which case the petals may be blotched or unblotched. They are almost always 

 protandrous, but homogamy may occur sporadically. In many cases the plant 

 is entirely dependent on cross-pollination. 



In the south Tyrol, Schulz found only genuinum, mostly with unblotched homo- 

 gamous flowers, presenting perfect or almost perfect radial symmetry. Elsewhere 

 he also saw the large-flowered form of this variety growing together with pimpinelli- 

 folium ; the flowers of both were visited by insects to the same extent. With a 

 pipette, Schulz carefully removed every trace of nectar from the nectaries of a number 

 of newly-opened flowers of pimpinellifolium, and smeared the glands with shellac. 



Fig. 70. Variations in the nectar-guides on the petals of Erodium cicutarium, L'Hcril. 



(From nature.) 



In spite of their nectar-guides, such flowers were visited by only a few insects, 

 while neighbouring blossoms both those of genuinum, and untouched ones of 

 pimpinellifolium received, just as before, a considerable number of visits. Here 

 again, the odour of the nectar serves as the chief means of attracting the insects, 

 the nectar-guides being of secondary importance in this respect, except when very 

 brightly coloured, and therefore conspicuous by contrast. 



At a time when unacquainted with Schulz's work, I arrived at similar con- 

 clusions while studying the variation of the nectar-guides of Erodium cicutarium 

 in the North Frisian Islands. The flower is here markedly protandrous, and always 

 bilaterally symmetrical, the upper petals being shorter, broader, and of deeper hue 

 than the lower ones. The nectar-guides are developed to a very varying extent; 

 in some flowers they have all but disappeared, in others they are very conspicuous. 

 But in all cases the insect visitors are the same, and blossoms which possess well- 

 marked nectar-guides are by no means more sought out than those almost lacking 

 them. Insects fly from one kind of flower to another without their selection being 

 influenced by the degree of prominence of the nectar-guides. This is noteworthy, it 

 being generally assumed that the guides indicate the position of the nectar to insects, 

 which in flowers without such markings, either do not find, or do not suspect the 



