7. GYNANDRÆ. Two isolated families, which probably have taken their origin from Liliifloræ, and have epigynous, mostly zygomorphic, and much reduced forms. Family 1. =Helobieæ.= To this family belong _only water- or marsh-plants_; _the endosperm is wanting_, and they possess an embryo with a very _large hypocotyl_ prolonged downwards and often club-like. The perianth is often differentiated into calyx and corolla; the flower is regular, and in the first orders to be considered, may be reduced to the ordinary Monocotyledonous type; there are, however, _usually found two_ 3-_merous whorls of carpels_ (Fig. 282), and thus in all 6 whorls, or again, the _number of carpels may be indefinite_; the number of stamens also may be increased, either by the division of the members of a whorl, or by the development of additional whorls. _Syncarps_,[26] with nut or follicular fruitlets, are _very common_, for example, in the first orders; in the last (Hydrocharitaceæ) the carpels are not only united, but the ovary is even inferior. The primitive type appears to be a hypogynous flower, similar to that of the Juncaginaceæ or Alismaceæ, with several 3-merous whorls, and free carpels, each with many ovules; the green perianth in this instance being no doubt older than the coloured ones. If we take a flower with this structure as the starting point, then the family developes partly into epigynous forms, partly into others which are so strongly reduced and exceptional that it is scarcely possible to refer them to the ordinary type. The family, through the peculiar _Zostereæ_, appears to approach the Araceæ, in which _Potamogetonaceæ_ and _Najadaceæ_ are included by some authorities. However, the inclusion of _Potamogeton_, and with it _Ruppia_ and _Zannichellia_, in the Juncaginaceæ appears quite correct. It would scarcely be right to separate _Zostereæ_ from these. Great stress has often been laid upon the similarity with the Ranunculaceæ which is found in the Alismaceæ, but it is scarcely more than an analogous resemblance.