{ 
3 
sa 
136 Rhodora [AuGusT 
winged, but they are wingless in Panctenis, A. pedicularia (L.) Raf. 
and A. pectinata (Nutt.) Pennell. But for proving kinship such 
lack of winging is negative. Moreover the seeds of Panctenis and of 
Agalinis are readily distinguishable; in Panctenis they are plumper 
and the reticulations of the seed-coat are slightly raised giving a sur- 
face roughened, minutely honeycombed; in Agalinis the thinner 
usually more angular seeds are practically smooth. 
Such is the correlation, not absolute, of characters by which to 
distinguish Aureolaria and Agalinis. Less may be urged for Otophylla. 
But certain features appear. Most important, I believe, is the reduc- 
tion of the anthers of the posterior pair of stamens. Correlated with 
this we find: stem retrorse-hispid (not ascending-scabrous to glabrous 
as in Agalinis); leaves broader, lanceolate to ovate, in one species 
usually incised at base, in the other bipinnatifid; pedicels shorter, 
so that flowers are nearly sessile; corolla (as in Agalinis linifolia) 
with no trace of yellow lines within; and seeds honeycombed. 
Now let us hold these three natural groups in one genus and attempt 
to find characters by which to distinguish the whole from allied genera. 
Allies are found in South America, and especially in Africa and Asia. 
(A) The open-throated corolla is satisfactory for our flora but can- 
not be used as against Sopubia, Graderia, etc. Moreover it is not 
constant within “Gerardia,” for in one section of Agalinis the corolla 
is progressively modified through stages present in various species ! 
into the flattened “two-lipped”’ corollas of A. divaricata (Chapm.) 
Pennell and A. filicaulis (Benth.) Pennell. 
(B) Equivalence of anther-sacs. In most of the allied genera we 
find a reduction of one sac of each anther.? But in certain of the Old 
World genera, and such New World allies as Afzelia (Seymeria) and 
Silvia this does not occur. Moreover absence of reduction is again 
purely a negative character. Otophylla is reducing its anthers by 
another process, also worthy of recognition. 
(C) Absence of bractlets on pedicel. Again a negative character; 
of. value in distinguishing from many Old World, but not from New 
World allies. 
(D) There is but one other point occurring to me, the only feature 
1 E. g. our common A. tenuifolia (Vahl) Raf. 
2 There is an American species, wrongly heretofore held in ‘‘Gerardia,’’ which shows this 
well, Gerardia hispidula Mart. This possesses other points of distinction, including bluish 
corollas and bractlets on pedicel. 
