104 BOTANICAL GAZETTE. 



The Anthers of Clethra.— To all eastern botanists our com- 

 mon Clethra, C. alnifolia, L., is certainly familiar, yet as far as I know 

 no one has noticed the striking peculiarities of the anthers. In the 

 more southern species, C. acuminata, Mx. , these are even more 

 marked. During the past summer I had opportunity to make a careful 

 study of both species in all stages except fruit. According to Benth- 

 am and Hooker (Gen Plant II, 603 j the genus Ch'thra is a waif as far 

 as the suborders of Ericacete are concerned — "genus anomalum," 

 they call it — but Dr. Gray (Syn. Fl. N. A. II, 17) places it among the 

 Pyrolincte. 



There is nothing peculiar in the development of the flowers. In 

 C. alnifolia there is a marked difference between the outer and inner 

 whorl of stamens while young, the latter being noticeably shorter. 

 From the very beginning the anthers are cxtrorse and when young 

 very decidedly epipetalous. They begin as mere knobs at the base of 

 the petals but soon become sharply sagittate, which shape they retain. 

 As long as the anthers are enclosed in the bud the filaments are bent 

 upon themselves, but differently in the two species. In C. alnifolia 

 they are shaped like a fish hook bearing the anther at the pomt cor- 

 responding to the barb, while the filament is represented by the shank. 

 In C. acuminata the filament resembles an interrogation point (?) ex- 

 cept that the first bend (counting from below upwards) isalmosta right 

 angle instead of a gradual curve. In all species these bends are more 

 or less marked as may be seen in various figures.* 



But in none of the figures referred to is to be seen anything pe- 

 culiar about the anther or connective, though several figure the en- 

 larged stamens. Not in any of the descriptions of foreign species 

 is there notice of anything in the structure of the anthers to provide for 

 their retroversion at anthesis common to the whole genus f 



As soon as the flower begins to open, the growth of the filaments, 

 which is extremely rapid at this time, pushes the anther beyond the 

 corolla, and it, relieved of the compression of the petals, immediately 

 spriags to a horizontal position. Tiie completion of obversion until 

 the anther becomes introrsc then proceeds more slowly though it is 

 accomplished in a few minutes after the first spring has taken place. 

 There seem to be two causes for this freak. The straightening of 

 the filament both lengthens the stamen and continues the somersault 

 begun by a special device, viz.: a cushion of turgid cells on the back 

 of the connective. This cushion is continuous with the filament, but 

 is not joined to the connective throughout its whole extent, being arched 

 away from it about che middle of the anther. The cushion divides 

 into two tongues, which taper to slender points as they pass down the 

 thecae. Under the microscope the cells of the upper part of the fila- 

 ment and of this cushion are seen to be turgid while the anther is 

 still held by the corolla, but soon after it escapes these become shriv- 

 eled and the cushion withers first. The outside cells are shown by a 



*Vi(le aieissn. in Mart. Fl. Bras. Vll, t. 64, 65, 66.— Lam. llhistr. t. 3li0.— Hot. Mag. 

 t. ]0.'-)7, 37-13 —Liiidl. Bot Keg-. 1842. t 23. 



-;-Vi(le DC. Piod. VII, 5-9. -Walp. Kep. II, 726; VI. 417; Ann. I, 479.— Mi(i. Fl. Ind. Uat. 

 II, Ki.ji..— Griseb. Fl. Bnt. VV. Ind. 141. 



