10 MEMOIRS FROM THE DEPARTMENT OF BOTANY OF COLUMBIA COLLEGE. 



are rarely auricled, and the blades often glandular-punctate. The blade is either 

 ()l)seurel}' or conspicuously nerved and rarely three-ribbed, and in one or two instances 

 it has two lateral impressions parallel with the midrib. Very rarely, too, we meet with 

 a keel on the under surface, as in the case of P. Paronychia. Towards and in the inflor- 

 escence the lea-\'es are sometimes reduced to foliaceous bracts, and all are subtended by 

 stipules in the form of sheaths or ocreae, to which they are articulated or with which 

 they are continuous. The stipules are of unusual form and characteristic. They are 

 united so that they form a sheath which surrounds the stem for a greater or less distance 

 above a node. These ocreae are either cylindric or funnelform ; the former usually hori- 

 zontally truncate at the summit and the latter oblique or two-parted. The summits may 

 be either naked or variously fringed with bristles, and in two subgenera a spreading, 

 collar-like rim is found. The ocreae are either glabrous or strigose, sometimes with 

 smocjth or ciliate ribs and rarely with a ring of more or less reflexed hairs around the base. 

 The inflorescence is both axillary and terminal, appearing as clusters, spikes, ra- 

 cemes or spicate racemes. These different forms may be solitary, geminate, or paniculate 

 and erect or drooping. Ocreae or more or less bract-like ocreolae subtend the flowers, 

 which are borne on usually fascicled, jointed, erect and stout or deflexed and often slender 

 pedicels. The subgenus Duravia is the only group in which the flowers are normally 

 solitary at the nodes and in which the inflorescence is what we may call spicate. There 

 are two interesting points suggested here. Fir§t, the phenomenon of cleistogamy, and 

 second, fertilization. Little has been written on the former subject. Flowers may be 

 produced on the roots as in the case of P. punctatum,^ or they occur within the ocreae in 

 certain species, as shown by P. Hydropiper^ and species of other subgenera.^ This latter 

 kind of cleistogamy appears in two different ways. In consequence of the morphology of 

 the plants of this genus, it is liable to occur in any species as illustrated in P. Hydropiper. 

 On the other hand, as in the case of members of subgenus Duravia, it is the normal con- 

 dition. Many interesting facts concerning fertilization have been recorded.* Both close 

 and cross-fertilization are common in the genus. The flowers of different species vary 

 in their structure and have adapted themselves as their environment directed. Some 

 have fragrant flowers with eight nectaries, situated at the base of the stamens, which se- 

 crete an abundant supply of honey. In other forms these organs, as well as the strength 

 of the fragrance, are less strongly developed, and so the scale descends until in such 

 plants as P. aviculare, P. HttoraJe and the like, the fragrance, showy calyx and nectaries 

 do not exist. The higher the coloring of the calyx the greater is the development of the 



1 Kearney, Coult. Bot. Gaz. 16 : 314. ^g. Coulter. Coult. Bot. Gaz. 17 : 91. 



"Meehan, Coult. Bot. Gaz 16 : 373 (erroneously published as P. acre). *H. Miiller, Fert. of Flowers, 509-516. 



