406 Dr. Planchon on Melianthese, 



compressed on its sides, and deeply cut into five unequal segments. Of these, 

 the two larger are in fact inferior, but in appearance superior in the expanded 

 flower ; then follow two lateral ones, narrower and shorter than the first ; 

 and lastly, on the posterior (in appearance anterior) angle of the base of the 

 flower appears the fifth segment, which, much shorter than the rest, is not 

 unlike a Phrygian cap, or rather a cowl, with its wide and oblique aperture 

 turned towards the inside of the flower, and its apex generally tapering into a 

 short pendent spur. Out of this concave sepal seem hardly to emerge four of 

 the comparatively small petals, which are inserted around a large horseshoe- 

 shaped concave gland filling almost all the upper (apparently lower) half of 

 the receptacle. Although the claws of those petals are all converging, and 

 even connected together towards their apex by means of the woolly indumen- 

 tum of their margins, their bases of insertion are however, according to rule, 

 alternate with the corresponding sepals ; that is to say, the two superior (ap- 

 parently inferior) petals are inserted each between the cowl-like sepal and the 

 adjacent lateral one ; the two other (lateral) petals each between one of the 

 lateral narrower sepals and the adjacent anterior (but apparently posterior) 

 one. In addition, an artificial expansion of those two inferior larger sepals will 

 show, alternating with them, a small imperfect fifth petal, generally reduced 

 to a thick linear unguis, with or without a rudiment of a lamina. The alter- 

 nation of the five petals with the calycine segments being thus successfully 

 ascertained, let us turn now to the more internal organs. The large gland 

 already noticed, and which is evidently an incomplete disc, fills a broad interval 

 between the two posterior (apparently anterior) petals and the two posterior 

 stamens. Hence the position of the fertile organs of the flower is remark- 

 ably eccentric, and their insertion nearer to the anterior sepals than to the 

 posterior cowl-shaped one. They consist of a four-celled ovarium and of four 

 slightly didynamous stamens, which, like the style, are gently turned back- 

 wards (apparently forwards in the expanded and inverted flower). Of these, 

 the two anterior (apparently posterior) are free and opposite to the larger 

 sepals ; while the small abortive petal alternates with them. The two others, 

 more or less connected at their base, are in fact, although less evidently, 

 alternate each with two of the connected petals. To arrive then at a perfect 

 symmetry of the flower one stamen only is required, and that, if developed. 



