144 



insertion, they are usually springing from the corolla, 

 sometimes however from the lengthened thalamus (fig. 8). 

 With respect to their shape they are as a rule normal, 

 once, however, I found the whole whorl affected in the 

 direction tending to petalody. Fig. 9 represents two of them ; 

 although the outlines of the anthers are recognisable, they 

 proved on examination perfectly foliaceous. 



The corolla is generally not longer than 4^ mM. and 

 not shorter than 3 k mM., consequently about half as 

 long as the fullgrown tubular florets. The mouth of the 

 shortened tube is wider. 



As has already been stated the lengthened thalamus 

 ends in a secundary headlet. The length of the peduncle 

 bearing the headlet, is, as shown in the figures (fig. 10 

 excepted, on which see infra) limited rather narrowly. The 

 headlet is — like a primary head — surrounded by a 

 séries of bracts and consists of tubular florets only, ail 

 of which are closed in the spécimens at my disposai. The 

 construction of the florets is normal, indications of a 

 second prolification were not to be seen. The secundary 

 headlets produced by tubular florets consist in fact of tubular 

 florets, nevertheless I found a small number, and thèse 

 close to the margin, which moreover possessed a few ligu- 

 late florets. 



Of the same spécimen I hâve also examined the ligulate 

 florets. Some of them proved to prolify. The whole appeared 

 to consist of a peduncled ligulate corolla the centre of 

 which bore a peduncled headlet. This secundary headlet 

 consisted of bracts and ligulate florets with (inferior) ovary 

 and two stigmas. Apparently thèse secundary florets were 

 fertile thus differing from the primary stérile florets. 



The second spécimen was much poorer of prolifications 

 and deserves on account of the slight différence from the 

 spécimen just described, no spécial description. 



The third spécimen, however, attracted attention at first 



