770 GALIACEyE. [Epigynous Exogexs. 



the first, third, and fifth leaves produce axillary buds, then in the whorl next above it, the 

 second, fom'th, and sixth leaves will probably be gemmiferous, according to the ordinary 

 laws of decussation. It is plamly impossible to say that what seem to be leaves are in 

 reality stipules, because they have no axillary buds ; for if that opinion were main- 

 tained, it would be necessai'y to assign the quality of stipules to a certam portion of the 

 leaves of such verticillate plants as Dysophylla stellata, in wliich only a part of the 

 whorls ever produces branches. 



2. If it is true that m Aspeinila two opposite leaves are frequently longer than the 

 others, that cu'cumstance may be reasonably ascribed to the greater development con- 

 sequent upon theu' higher functions, and to their peculiar position on the stem ; and it 

 is equally time that m the greater part of SteUates no trace whatever of any kind of 

 difference between the leaves can be detected, as is most remarkably the case in those 

 smTOunding the flowers of Crucian ella maritima. 



3. The argument derived from the occasional coimection of the leaves by a membrane 

 can hardly be allowed much weight, when it is remembered that in such cases the inter- 

 mediate leaves are less like stipules than in those cases where no membrane exists ; 

 compare Asperula cynanchica, or littoralis, or longiflora, with such genuine Cnicianellas 

 as C. maritima. 



4. The comparison of the supposed stipules of Stellates and the setae of Sperma- 

 cocese is inadmissible, because the former are at all events single simple organs, be 

 they what they may, while the setae of Spermacoceae are the result of the splitting of 

 two parallel- veined stipules, and therefore will necessarily be uncei'tain in number. 



These arguments do not, however, by any means exhaust the question ; and therefore 

 I proceed to make a few additional remarks upon a point not yet adverted to. It is in 

 Aspeinila, more than in any other genus of the Order, that is to be found e-s-idence 

 favoui'able to the supposition of M. De CandoUe and his followers. In A. longifloi^, 

 cjaianchica, and some others, the lower whorls are in the usual state, but the upper 

 ones are reduced to two perfect leaves, with one or sometimes two teeth or subulate 

 processes between them, which remain. In this condition the structure of Asperula is 

 so very hke that of many Spermacoceous plants, that the analogy between them seems 

 indisputable ; aud I presume that it was such cases which first led to the theory under 

 consideration. 



It is, however, to be remembered, that in Stellates the supposed stipules are always 

 what first disappear in the process of reduction in the number of foliaceous appendages ; 

 but that in Cinchonads it is in many cases the leaves which are first lost when such a 

 reduction takes place. The latter fact is readily verified upon reference to any of the 

 capitate Spermacoces, where the bracts are eWdently stipules, and especially to S. 

 calyptera, in which the leaves are gradually merged m the large membranous cup that 

 subtends the flowers, while the stipules suffer no diminution. The same circumstance 

 may be observed in several Brazilian Cinchonads allied to Psychotria barbiflora, and 

 in Pfederia foetida. It is also possible that the large colom'ed involucrura of Cephaelis 

 is, at least in some cases, formed by the excessive development of stipules and suppres- 

 sion of the leaves, for such is undoubtedly the case in a Sierra Leone plant in my pos- 

 session, which I presume is the little-knowTi C. bidentata of Thmiberg. These facts 

 render it more probable than ever that Stellates and Cinchonads are essentially dif- 

 ferent Natural Orders ; for they would seem to show that while the first has verticil- 

 late fohaceous organs, the most imperfect of which have the greater tendency to 

 disappear, the second has verticillate fohaceous organs, the most perfect of which have 

 the greater tendency to become abortive. I need scarcely add, that after a full con- 

 sideration of this point I retain my original con\'iction, that the apparent leaves of 

 Stellates are really leaves, and not stipules, and that the Order is as distinct from 

 Cinchonads as Nightshades from Figworts, Verbenes from Labiates, and I might even 

 add, as Cinchonads themselves from Umbellifers. — See Bot. Reg. 1838. 55. To be 

 consistent, then, we must either combine Caprifoils ^\•ith Cmchonads, or we must pre- 

 serve Stellates separate. Properly speaking, the appellation Rubiaceae should be 

 confined to the latter gi'oup, as it comprehends the genus Rubia ; but that name has 

 been so generally applied to the larger mass now comprehended under the name of 

 Cinchonads, that I find it better to abolish that of Rubiaceae altogether. 



Natives of the northern parts of the northern hemisphere, where they are extremely 

 common weeds, and of high mountainous regions in Peru, Chih, and Australasia. 



First among them stands Madder, the root of Rubia tinctoria, one of the most import- 

 ant dyes \\ith which we are acquainted ; a quahty in which other species of Stellates 

 participate m a greater or less degree. The roots of Rubia cordifoHa (Mmijista, Roxh.) 

 yield the Madder of Bengal, and fonn even an article of the export commerce to 

 Em'ope, under the name of Munjeeth. Rubia angustissima, from Tong Dong, has also 

 highly-colom-ed roots, and Rubia Relboun is the Madder of Chih. It has been remarked 



