126 THE JOURNAL OF BOTANY 



Swynnerton's specimens of this Gazaland tree it is found that the 

 racemes on what are male trees bear male flowers only. But in 

 the case of the racemes from what are female trees it is found 

 that, while most of these bear female flowers only, some bear here 

 and there male flower-buds associated with female flowers in the 

 axils of the same bracts. These male flowers on female racemes 

 are rather smaller in size than the male buds on male racemes, 

 with the calyx sometimes 4-lobed in place of 2-lobed ; they are 

 also rather later of development than the female flowers with 

 which they are associated. In the Gazaland plant we find, there- 

 fore, a repetition of the conditions manifest in the Comoro 

 species. The only difference is that, whereas Swynnerton's 

 material fully explains the floral characters of the Gazaland plant, 

 Boivin's specimens of the Comoro one only exemplified that con- 

 dition of its female in which males are associated with female 

 flowers, while Humblot's specimens of the same species exem- 

 plified, along with strictly male individuals, only those females in 

 which no males accompany the female flowers. It is noticeable 

 that, just as in the Gazaland plant, the calyx in male flowers on 

 female racemes may be 4-lobed whereas that in male flowers on 

 male racemes is 2-lobed, so in the Comoro plant the calyx in male 

 flowers on female racemes may be 5-lobed whereas that in male 

 flowers on male racemes is 2-lobed. It is interesting and im- 

 portant to observe that, although Baillon and Pax have been ahke 

 handicapped, owing to the absence of complete material, these 

 able botanists have independently recognized the true affinity of 

 the species. 



The presence in Gazaland of the interesting tree discovered by 

 Swynnerton has incidentally led to another conclusion. In 1909, 

 Pax, in dealing with two plants collected in German East Africa 

 by Goetze and Hoist respectively, based upon these a new and 

 apparently very distinct genus, Holstia Pax (Engl. Bot. Jahrb. xhii. 

 220). With this genus Holstia the Gazaland plant agrees in 

 having terminal inflorescences ; in having a ring of disk-glands in 

 the male flower and a disc in the female flower ; in show^ing the 

 same relative position of petals, disk-glands, and outer stamens ; 

 and in having the same ovary and stigmas. The Gazaland plant 

 difters from the only Holstia in which the female flower is known 

 in having the female calyx wath short triangular lobes and the 

 female petals much longer than the calyx, whereas in Holstia the 

 female calyx is deeply partite with spathulate lobes, and the 

 female petals are shorter than the calyx. But these differences 

 are at best only of sectional value, and leave no doubt that the 

 Gazaland plant is congeneric with the species upon which Holstia 

 was based. Since, however, the Gazaland plant is at the same 

 time certainly a Tannodia, the reduction of Holstia to Baillon's 

 genus is unavoidable. 



The specific descriptions which Baillon, Pax, and Moore have 

 supplied are so precise that nothing would be gained by their 

 repetition. Having regard, however, to the difficulties which 

 have been experienced in referring the various species of Tannodia 



