264 



CONTRIBUTIONS FROM THE NATIONAL HERBARIUM. 



side and on the upper part of the inside, divided at the tip into 4 or 5 rounded 

 lobules. Ovary ovate, only partly adhering to the perianth; ovule and style 

 inserted laterally. Style 3 to 3.5 nun. long, slender, sparsely hairy; stigmas 

 rather short, twisted. 



Receptacles of the infructescence very shallow. Achenia not fleshy nor con- 

 fluent, pubescent, carinate on the suture of the lobes of the perianth. Seeds 

 small (about 8 mm. long, (! nun. in diameter), ovate, slightly compressed. 



Costa Rica: Quebrada de I'otrero Grande near Terraba, Diquis Valley, at 

 an altitude of 150 meters, Pittier, flowers and fruit, March IS. 1898 (Instltuto 

 ffs.-geog. Costa Rica no. 12051). 



Explanation of Plate 2.'!. — a, h, port ions of ;i branch bearing male inflorescences; c, 

 piece of bark from the same to show the strigose hairs; </, calyptrate bract (or bractsi 

 which shields 1 he male inflorescence; e, section of :\ young male inflorescence; /. an 

 involucral scale; {j, bracteoles between the male flowers; h, a male flower; i, infructescence 

 seen from below; j, the same from above; /,-, a section through a portion of the same, 

 showing that the carpels (pistils) are completely immersed; 1, a single pistil (fruit I ; 

 m, seeds of different shapes; n, embryo; o, portion of one cotyledon and axis. Reduced 

 one-half from Hooker's Jcones, loc. cit., pi. LMl.'iT. In tin* original all the figures more 

 or less enlarged except a, ft, i, o. The explanations are as in llemsley ; for exclusions see 

 observations immediately below. 



Fia. 40. — Floral details of Castilhi foliar, a, Calyptrate bract protecting the male in- 

 Jlorescencc ; b, forms of stamens; c, female flower; d, infructescence. a, After llemsley ; 

 d, after Warburg, <r, d, Natural size ; 1>, c, scale II. 



According to observations made by me and published in 11)03," the leaves of 

 seedlings have remote marginal teeth that soon disappear, except ;it the base 

 of the blade. On dried specimens from mature trees no such detail is notice- 

 able, which shows that the teeth tire really tufts of hairs. Another interesting 

 feature, brought to light by Mr. Hemsley, is the calyptra-like bract that covers 

 the male floral buds. (Fig. 40, a.) In plate 2651 of Hooker's Icones, already 

 referred to, we have excellent representations of parts of ('. fallax in the draw- 

 ing of the leaf and in figures 1 to 7. Figure 8 shows the anthers as hairy, which 

 is not the cast 1 with this species, and the stamens are not regularly geminate. 

 The drawings of fruits (figures <) to II) do not belong to C. fallax, but more 

 likely to C, guatemalensis. Figure 12 can hardly refer to a Castilla, unless it 

 represents, instead of a pistil, a closed female flower. Of figures 13 to 15 it may 

 be said that they apply to any of the small-seeded forms of the genus. 



As to the further distribution of C. fallax, we are reduced to oral infor- 

 mation, not always from very reliable sources. Personally, the writer has met 

 the tree all over the Diquis Basin, at altitudes not exceeding GOO meters, and 

 also near the mouth of the Savegre River, and along the Goto River in the 



"Boletin del Institute fisico-gcografieo, vol. 3, p. 129 (1!>03), 



