NORTH AMERICAN GERANIACEAE. 79 



decumbent forms with this simple type, for their leaves are scattered and, in G. pusillum, 

 etc., the 2-flowered peduncles arc frequently opposite the leaves. But in G. Sihiricum, 

 which is more or less regularly dichotomous, tlie forks are occupied by mostly 1-flowered 

 peduncles, clearly the terminal shoots, and in the slender-stemmed species with the pe- 

 duncles opposite the leaves the former are as certainly terminal, the leafy prolongation 

 of the stem being in reality an axillary branch. (Note 2.) 



The most interesting part of a biological study of Ger-anium is that relating to its 

 pollination, and in this respect it has a historical iuterest, for it was the detection of tufts 

 of hairs friugiug the bases of the petals in G. sylvaticam, which led Sprengel just a 

 century ago to examine the flowers closely in an eff'ort, if possible, to determine their 

 use, for he was utilitarian enough to believe that not even a hair grew without purpose. 

 His study, cari-yiiig him farther than he had at first anticipated, led to the publication, in 

 1793, of a work wliich he quaintly styles the "Disclosed Secret of jSTature," which not 

 only contains much of scientific value, but is one of the most intei-esting biological 

 treatises ever written, and, guiding the researches of Darwin, Miiller, Delpino, and many 

 other observers, has contributed not a little to the foundation on which the theory of 

 evolution by natural selection i-ests. 



ISTo great attention has been given to the pollination of our American species; but, so 

 far as they have been observed, the facts agree closely with those brought out by Miiller's 

 study of the same or related species in Germany, so I cannot do better than refer to his 

 admirable account of the latter.^ I may add that the larger-flowered species are conspic- 

 uously protandrous, their two sets of stamens becoming erect, and dehiscing successively 

 at the centre of the flower, which, after their anthers have fallen, is occupied by the now 

 expanded stigmas; while in the species with smaller, less conspicuous flowers, the dura- 

 tion of the staminate stage is much shortened, or, in G. j^usillum, where the number of 

 stamens with anthers is reduced one-half, the flowers are synacmic or slightly proto- 

 gynous and certain of self-pollination if crossing is not secured. 



The large-flowered species are, in the main, incapable of self-pollination, and depend 

 exclusively (except in occasional synacmic flowers) upon the good ofiices of insects, 

 chiefly bees of diflerent kinds, which are attracted by the conspicuous petals, the veins 

 of which point to an abundance of nectar, secreted by five prominent glands at the bases 

 of the sepals. This is protected from inclement weather by tufts of bail's fringing the 

 petals below, and, usually, from creeping insects like ants which cannot effectively bring 

 about cross-fertilization, by the retrorse or glandular pubescence of the pedicels or stem. 

 Doubtless the facility with which several of the small-flowered species gain a foothold 

 in new countries is to be explained by their ability to self-fertilize where they are not 

 visited by appropriate insects, as well as by their annual habit and abundant seeding. 

 G- Bohertianum alone, with nar-rower flowers contracted into a sort of tube, is adajited 

 to pollination by long-tongued insects like the Si/rjyJndae, which visit it fi-eely in Eu- 

 rope, though I do not know that its American visitors have been recorded. (Note 3.) 



The ripening carpels, as they dry, contract in such a manner that the outside is short- 

 est, so that there is a tendency for their ends to bend outwardly; and, ultimately, after 

 dehiscing along the ventral suture, they break away at the base and suddenly curve up- 



' Fei'tilization of Flowers, p. HO. 



MEMOIRS BOSTON SOC. NAT. HIST., Vol.. IV. 12 



