All botanists acquainted with the family Rubiaceae will agree that the present subdivision is far from satisfactory and that more than one of its tribes are either artificial or ill-defined or both. The genera dealt with in this paper are said to belong to the Mussaendeae, but the distinction between this tribe and the Hedyotideae as defined by BENTHAM and Hooker f. (Oldenlandieae K. SCh.) rests merely on the succulence or non-succulence of the fruit and must therefore be regarded as both artificial and ill-defined: artificial, because from a morphological point of view the difference between dry and fleshy fruits is certainly not more important than that between the capsular and schizococcous fruits brought together in the first group and not more weighty than that between the various kinds of berries and drupes referred to the second; ill-defined, because the baccate fruits are sometimes dehiscent and the schizococcous ones more or less fleshy. The absence of a sharp line of demarcation separating the dry from the fleshy fruits doubtless explains the fact that the distinction has never been rigorously applied: Mussaenda L., the standard genus of the tribe with fleshy fruits, at present comprises several species provided with capsules, and plants with drupaceous fruits, by BLUME rightly referred to a genus of their own, Metabolos, have been included by BENTHAM and Hooker f. in Hedyotis L. and by K. SCHUMANN in Oldenlandia L. RIDLEY’S genus Pomazota was referred to the Hedyotideae, because the fruit, though soft and succulent, opens at last, but it is, as I will show elsewhere, identical with Coptophyllum KORTH. non GARDN., which on account of its baccate fruit was put in the Mussaendeae. Other examples might be adduced, but these will suffice to show that the distinction is a source of confusion and should be given up as soon as possible.