140 Mr. Brom'n's Observations on the 



This irregular expansion in Boopis, which renders even the ge- 

 neric name improper, and at present the want of satisfactory cha- 

 racters to distinguish it from Calycera, are objections to the name 

 M. Cassini has chosen for this family ; while thatof Cfl/^ce?'e(«, which 

 I have proposed, derived from the genus first described, and appli- 

 cable to all the genera of the order, appears to me unexception- 

 able : especially as there seems no reason to doubt that the part 

 which I have considered as calyx in Boopidece is really such ; its 

 divisions being generally in equal number, and alternating with 

 those of the corolla. It may be observed that a like alternation 

 of the divisions of the pappus with the segments of the corolla 

 obtains in those genera of Compositae where both parts are in 

 equal number. But in some cases, where the division of pappus 

 is still further reduced, the same alternation does not exist, espe- 

 cially in those genera having vertically compressed pericarpia and 

 two arista;, as Spilanihus and Sabnea. 



The absence of " discus epigynus" in Boopidea is a necessary 

 consequence of the accretion of the base of the style with the tube 

 of the corolla. It seems to me, however, that a modification of the 

 same organ may be traced in the five thickened areola observa- 

 ble within and near the base of the tube formed by the filaments 

 in Acicarpha spathulata; and much more distinctly in the same 

 situation in Boopis balsamitifolia, where they have the appearance 

 of five adnate fleshy bodies alternating with the filaments. 



This apparent decomposition of the glandular disk in Boopidea, 

 compared with its state in Composita;, as well as its transposition 

 and the alternation of its parts with the stamina, seem to give 

 some additional support to the conjecture I have formerly ha- 

 jiardtd in the paper on Proteacece, published in the Society's Trans- 

 iictions (vol. X. p. 159) ; namely, that in several families — for the 



hypothesis 



