There can be little doubt as to a very close relationship between _Sigillaria_, _Lepidodendron_, and _Bothrodendron_. _Sigillaria_ seems to have outlived _Lepidodendron_ and _Bothrodendron_. The two latter genera are recorded from Upper Devonian rocks in several localities, _Bothrodendron_ being particularly abundant in the pre-Carboniferous floras of Bear Island and other parts of the world. A remarkable stem described by Dr White[659] as _Archaeosigillaria primaeva_ from Upper Devonian shales of New York is spoken of by him as “one of the most highly developed representatives of a fairly distinct archaic group foreshadowing the later genera _Bothrodendron_, _Sigillaria_, _Lepidodendron_ and _Lepidophloios_.” The type-specimen, when first discovered, consisted of an apparently unbranched stem reaching a length of 5 metres. From the swollen basal part Stigmaria-like rootlets spread into the surrounding shale. At a higher level the fissured bark shows indistinctly defined leaf-cushions which pass gradually upwards into cushions and scars arranged in closer order on regular vertical ribs. The surface-features in this region are practically those of a ribbed _Sigillaria_. Traced farther upwards the vertical ribs die out and cushions of the Lepidodendroid form cover the surface of the bark. The leaf-scars, with a supraposed ligular pit and two vertically elongated parichnos-scars, are said to bear a closer resemblance to those of _Sigillaria_ and _Bothrodendron_ than to the leaf-areas of _Lepidodendron_. Nothing is known as to the anatomy of this stem, nor have fertile shoots been discovered. In the absence of more trustworthy evidence than is available conclusions of a phylogenetic nature must be accepted at their true value. It is however legitimate to describe _Archaeosigillaria primaeva_ as one of the oldest examples of a lycopodiaceous plant which shows well-preserved external features, and these are of exceptional interest as indicating a combination of generic characters. This Devonian type lends support to the view that _Lepidodendron_ and _Sigillaria_ are offshoots, differing from one another in comparatively unimportant points, from a common ancestral type. The generally accepted statement that arborescent Palaeozoic Lycopodiales bore their sporangia on specially modified leaves (sporophylls) grouped in cones which were usually produced at the tip of slender branches, has recently shared the fate of most rules. Prof. Bower in his _Origin of a Land Flora_ mentions a Belgian specimen of _Pinakodendron musivum_ Weiss from the Westphalian series (Middle Coal-Measures), to be described by Dr Kidston, which bore its sporangia “associated with the leaves of certain portions of the stem, without any cone-formation. The fertile and sterile portions are distinguished only by the presence or absence of sporangia[660].”