432 PROCEEDINGS OF TIIE AMEKICAN ACADEMY. 



Eupatorium tribe needs restatement with greater detail, namely the 

 apical appendage of the anthers. It has long been customary to 

 separate the subtri])e Piqucrinae on the ground that this apical ap- 

 pendage was there absent, yet to this subtribe \\ix\Q been referred 

 several genera, such as Adenostemma, Gymnocoronis, Ilarfw right in, 

 Podophania, Decachacfa, Hdogyne, and Eupntoriopsi.s, which exhibit 

 such an apical appendage in various degrees of development, while \n 

 Enpatoriinii proper there are several species in which the appendage 

 is decidedly rudimentary. It will thus be seen that the distincti(m 

 between the Piquerinae and the Agcratinae breaks down completely 

 unless it can be supplemented by other characters or restated with 

 proper qualification. Persistent search has failed to discover any 

 correlated differences, so it is necessary to make the most of the unthcr- 

 appendage. 



Examination of the typical Piqiicrinae, such as I'iqiicria, Opliri/os- 

 poru.s-, etc., shows that the anthers are entirely destitute of apic-al 

 ajipendage, nor is the connective upwardly thickened or expanded in 

 a way to cover the cells. If, however, the (loul)tful genera above 

 mentioned are examined, it becomes evident that the rudimentary 

 appendage assumes several distinguishable forms. In Drcachacia 

 and Podopluinia it consists merely of an exceedingly short cushion-like 

 expansion of the dilated summit of the connective. The connective 

 is tlius without change of texture broadened out until it partially or 

 entirely covers the apices of the locidi, the whole anther being de- 

 cidedly blunt or even slightly refuse. It is to be noted that there is 

 here no membranaceous appendage in the stricter sense, and it seems 

 best to retain these genera in the Piqiwrinae. 



The next stage in the development of the apical appendage is seen 

 in the three genera Adenostemma, Gymnocoronis, and llartwrightia. 

 In these there appears always to be at least a very short and decidedly 

 retuse membranous appendage. This may be a single or simple 

 structure or in Adenostemma it is sometimes so deeply divided in the 

 middle that it virtually becomes two small membranous tips crowning 

 and slightly prolonging the two locidi. In other respects these three 

 genera manifest striking similarities, such as the form of the corolla 

 and especially in the uniforndy glandular faces of the achenes. Both 

 on account of the seemingly close relationship among themselves antl 

 in order to permit a more precise definition of the Piquerinae it seems 

 desirable to class these three genera in a new co-ordinate subtribe, the 

 Adenostemmatinae . 



In Helogyne and Eupatoriopsis the apical appendage assumes the 



