NORTH AMERICAN GERANIACEAE. 93 



ti-unks, of which 0. articnlata^ is a representative. On the other hand, there is no doubt 

 that the latter approach the bnlbiferoiis tj])e, through the enneaphylla form already de- 

 scribed. 



Commonly the leaflets are digitately clustered at the end of an elongated petiole ; biit 

 the pinnate type, represented among our species by O. Berlandieri, is not uncommon in 

 the tropical American species, and in the section Biopliytmn the elongated leaves consist 

 of a large number of crowded leaflets. On the other hand the suppression of all but the 

 terminal leaflet results in the iniifoliolate leaf of 0. dichondraefolia and a few species 

 that do not occur in our teri-itory; and in O. rusciformis (sometimes quoted and figured 

 as O.frutescens), the reduction and final abortion of the leaflets, accompanied by an lui- 

 usual dilatation of the petioles, results in as perfect phyllodia as those of Acacia, — but 

 these stand with their edges tangential to the stem, not radial as in the latter genus. A 

 transition to this type is afforded by a considerable series of species in which the petioles 

 are more or less dilated and efficient in assimilation. On the other hand, Hildebrand^ 

 has jiointed out that the leaves of seedlings of O. Lasiandra are at first trifoliolate, 

 though the later leaves consist of 7-9 leaflets. 



The leaves of Oxalideae have long been IvUOAvn to possess the power of movement in 

 a marked degree. The change of position of the leaflets at night, so that they are then 

 exposed to a minimum cooling by radiation — the so-called sleep of the leaves — is a fa- 

 miliar example. Under the influence of light and warmth, this position is exchanged in 

 the daytime for one in which they have the best exposure for assimilation; but a sudden 

 and intense illumination causes the leaflets to close again more or less, and continued ex- 

 posure to either light or darlvuess for a period of days breaks the I'eadiness of their re- 

 sponse to either, while it has been shown that if so fastened that they cannot close at 

 night, they soon become feeble or die. 



In Amrrlioa, the leaflets are usually in evident motion, from an intensification of their 

 circumnutation, — a movement which has been carefully studied by Mr. Darwin in many 

 plants. In this respect they resemble the small lateral leaflets of Desmodium gyrans, the 

 common " telegraph plant " of greenhouses. It is also known that the leaves of Bio- 

 jyhyftwi, to a certain exteut those of Oxalis acetosella, and in a still less degi-ee of other 

 species, are influenced by shocks, which cause the leaflets to assume the nocturual posi- 

 tion with more or less rapidity, as in Mimosa, Rohinia, etc. The seat of these different 

 movements is in the pulvini near the base of the general petiole and of the short stalks 

 of the leaflets, and in the midribs of the latter. 



The flowers of most species are solitary or uuibellately clustered; but in the former 

 case the peduncle is 2-bracted some distance below the flower, and the umbels are 

 likewise subtended by two or more bracts. In specimens of O. violacea ^ some of the 

 branches of the umbel have been observed to branch, and this, with the constant presence of 

 the bracts referred to, indicates that the inflorescence in these cases is in reality a reduced 

 cyme. A t3q)ical dichotomous cyme is, in fjict, regularly developed in luxuriant speci- 

 mens of O. corniculata, var. sfricta^. In 0. Brasiliensis, bulblets have been observed, 



'Bot. Magazine, ex, PI. G748. * On this subject see Zuccarini : Nachtrag Monogr. Amer. 



-Bot. Zeitmig, 1887, 3. Oxalisarten, 100; WyiUer : I'riugslieira's Jalirb. vviss. Bot. 



^Zuccarini: Monogi'. Oxalid., 19; .Taiquiii : ITort. ix ; Eichli r : BUitheuiliagrainme, 11, 80+, etc. 

 Viiulob. 84, ri. 180. 



