THE STAMEN 113 



The terms winged and flattened, as commonly applied descriptively to 

 filaments, imply a modification of a terete form, but winged filaments 

 are usually more primitive than the terete type, rather than more 

 advanced. The broad, short filament in members of some primitive 

 families — Magnoliaceae (Fig. 51), Annonaceae — is hardly more than a 

 narrowed laminar base. The long, slender, terete filament is a highly spe- 

 cialized type. 



Though typically ephemeral, the filament is sometimes fleshy and long 

 persistent, or woody and retained even on the fruit — Kingia. It may be 

 much enlarged and may have appendages — at the top, Mahonia; at the 

 base, Violu. The morphological nature of the appendages is difficult to 

 determine and is clearly various. Some appendages are glandular ex- 

 crescences or modifications of form correlated with pollination methods. 

 The presence, type, and origin of vascular tissue in these appendages 

 is of much value in interpreting their nature. The "hood" and "horn" 

 of the complex stamen of the Asclepiadaceae have been called append- 

 ages of the stamen, but only the horn is an appendage, as shown by 

 ontogeny and anatomy; the hood is formed by the filament itself, re- 

 flexed upon its base. The petaloid "appendages" of the stamens of 

 Potamogeton are adnate bracts. Stipulelike appendages at the base of 

 the filament are occasional and characteristic of some taxa. They appar- 

 ently vary in morphology; some seem clearly stipular; others represent 

 vestiges of lost stamens of an ancestral fascicle, as in Parnassia, Sassafras, 

 Benzoin; still others are merely glandular proliferations. Lateral ap- 

 pendages may be prominent, as in some Amaryllidaceae, where they 

 become petaloid and connate, forming a corona. Anatomical structure 

 and comparison with related taxa are necessary to determine the nature 

 of these appendages. 



Anther-filament Relationship. A considerable number of descriptive 

 terms have been applied to types of attachment of anther to filament, 

 but these are conflicting and often obscure in meaning. Morphologically, 

 the anther is not "attached" to the filament, because both are parts of 

 one organ, but the terms are of descriptive and taxonomic value. In the 

 simpler type of stamen, without sharp distinction of anther, where the 

 filament is clearly continuous between the sporangia, the anther is 

 called innate; in similar stamens, where the connective is less prominent 

 and differs from the filament tip, the anther is basifixed. Where the 

 anther appears to be attached laterally to the distal part of the filament, 

 the anther is called adnate, in contrast with innate. Adnate is an un- 

 fortunate term here and should be dropped, because it implies, incor- 

 rectly, a lateral fusion of anther and filament. There is, rarely, a true 

 adnate condition, as in the Melastomaceae and Ericales, where the 

 more or less pendent and inverted anther has become fused to the 



