324 THE JOURNAL OF BOTANY 



(6) Cliili and ArgeJitina concern us but little. This area is 

 remarka])le chiefly for the Cruckshanksiea?, which are endemic, 

 and the curious, pustulate, toothed-leaved Heterojjlii/lUBa, from 

 Northern La Plata, also endemic. For the rest, the only Rubia- 

 ceous genera this area shares with the rest of w^armer America 

 are the ubiquitous Oldenlandia, Nertera, and certain Galieae. 



"i= * * * 



It is rarely a matter of much difficulty to recognize, almost at 

 sight, the family characters of a Eubiaceous plant, even in the 

 dried state. The leaves are invariably opposite or verticillate, and 

 simple, with perfectly entire margins. '= Between the petiole-bases 

 of the two leaves, or of each adjacent, pair of leaves at a node, an 

 " interpetiolar " stipule appears in most cases. This may be rapidly 

 caducous, but it can be detected in association with the youngest 

 leaf-members ; and even when the stipule has fallen, its trace is 

 present in the form of a scar or line. These characters, coupled 

 with the presence of an epigynous isomerous flower, establish any 

 specimen possessing them as a Rubiacea beyond doubt; with the 

 practical exception of the familiar tribe Galieae, including some 

 10 per cent, of the family, in which the foliar organs are arranged 

 in whorls of two to eight or more at each node. These may be 

 brought into line with the rest of the family by supposing that 

 some of the "leaves" at each node are modified interpetiolar 

 stipules. 



The morphology of the foliar organs in Galieae is discussed at 

 length in a masterly manner by Lindley in his Vegetable Kiyigdom 

 (ed. 3, pp. 768-771). This author separates this tribe from the 

 rest, according it family-rank — Galiaceae. I strongly favour the 

 same course. 



An interesting transition between this last-named group and 

 theRubiacece proper is furnished by the Mexican Didymcea (fig. 3). 

 In this case the stipules are appreciably different from the leaves ; 

 each consists of a bipartite, glandular, erect structure which at 

 length becomes recurved and functions as a double-hooked climbing 

 organ. The same transition, indeed, is reflected in the divided or 

 fimbriate stipules in other Rubiaceous tribes, notably the herbaceous 

 S])ermacocece. 



The theoretical principle of these interpetiolar stipules is 

 readily explicable by a diagram. 



The accompanying figures, 1, 2, represent a portion of stem 

 bearing two opposite leaves. A, B, at a node. Each leaf (fig. 1) 

 has its pair of stipules at the petiole-base, A with stipules a^, a^, 

 B with by, &2- I^ it be conceived that stipule a^ of leaf A travels 

 around the stem to fuse with stipule b-^ of leaf B, as indicated by the 

 dotted lines and arrows in fig. 1, while stipule a.^ behaves similarly in 

 regard to stipule 5^„ the condition indicated in fig. 2 will be arrived 

 at. The stem-axis will bear interpetiolar stipules composed of the 



* The sole exceptions to this latter are afforded, in the American genera, 

 at any rate, by the genus Pentagonia, some species of which have pinnately- 

 lobed leaves, and Heterophyllcea, a monotypic genus with toothed leaves. 



