Sept. 20, 1883] 



NATURE 



507 



from which aerial steins are sent upwards. I am satisfied that 

 there is not a shadow of foundation for such a belief. The same 

 authors, along with their distinguished countryman, the Marquis 

 de Saporta, believe with Brongniart that it is possible to separate 

 Sigillaria widely from Lepidodendron. They leave the latter 

 plant amongst the Lycopods, and elevate the former to the rank 

 of a Gyrunospermous exogen. I have in vain demonstrated the 

 existence of a large series of specimens of the same species of 

 plant, young states of which display all the essential features of 

 structure which they believe to characterise Lcpidodmdron, \\ hilst, 

 in its progress to maturity, every stage in the development of the 

 secondary wood, regarded by them as characteristic of a Sigil- 

 laria, can be followed step by step. 1 Nay, more : my cabinet 

 contains specimens of young dichotomously branching twigs, on 

 which one of the two diverging branches has only the centripetal 

 cylinder of the Lepidodendron, whilst the other has begun to 

 develop the secondary wood of the Sigillaria.* 



The distinguished botanic of the Institut, Ph. van Tieghem, 

 has recently paid some attention to the conclusions adopted by 

 his three countrymen in this controversy, and has made an 

 important advance upon those conclusions, in what I believe to 

 be the right direction. He recognises the Lycopodiice ons 

 character of the Sigillaria, and their close relations to the 

 Lepidodendra ; 3 and he also accepts my demonstration of the 

 unipolar, and consequently Lycopodiaceous, character of the 

 fibro-vascular bundle of the Stigmarian rootlet, a peculiarity of 

 >tructure of which M. Renault has hitherto denied the existence. 

 But along with these recognitions of the accuracy of my con- 

 clusions he gives fresh currency to several of the old errors 

 relating to parts of the subject to which he has not yet given 

 personal attention. Thus he considers that the Sigillaria, though 

 closely allied to the Lepidodendra, are distinguished from them 

 by possessing the pawer of developing the centrifugal or exogen- 

 ous zone of vascular ti>sue already referred to. He characterises 

 the Lepidodendra as having "unseul bois centripete," notwith- 

 standing the absolute demonstrations to the contrary contained in 

 my "Memoir " xi. Dealing with the root of Sigillaria, which in 

 Great Britain at least is the well-known Stigmaria ficaides, fol- 

 lowing Renault, lie designates it a "rhizome," limiting the term 

 root to what we designate the rootlets. He says, "Le rhizome 

 des Sigillaires a la meine structure que la tige aerienne, avec des 

 bois primaires tanti'l isole; a la peripheric de la moelle, tantot 

 confluents au centre et en un axe plein ; seulement les fasceaux 

 libero-ligneux secondares y sont separes par de plus la r ges 

 rayoDS," &c. 



Now, Stigmaria being a root, and not a rhizome, contains no 

 representative of the primary wood of the stem. This latter is, 

 as even M. Brongniart so correctly pointed out long ago, the 

 representative of the medullary sheath, and the fibro-vascular 

 bundles which it gives off are all foliar ones, as is the case with 

 the bundles given off by this sheath in all exogenous plants. But 

 in the Lepidodendra and Sigillaria, as in all living exogens, it is 

 not prolonged into the root. In the latter, as might be expected 

 a priori, we only find the secondary or exogenous vascular zone. 

 Having probably the largest collection of sections of Stigmaria 

 in the world, I speak unhesitatingly on these points. M. van 

 Tieghem further says, ' La tige aerienne part d'un rhizome 

 rameux tres-developpe nomme' Stigmaria, sur lequel s'inserent a 

 la fois de petites feuilles et des racines parfois dichotomies.' I 

 have yet to see a solitary fact justifying the statement that leaves 

 are intermingled with the rootlets of Stigmaria. The statement 

 rests upon an entire misinterpretation of sections of the fibro- 

 vascular bundles supplying those rootlets and an ignorance of the 

 nature and positions of the rootlets themselves. More than 

 forty years have elapsed since John Eddowes Bowman first 

 demonstrated that the Stigmaria; were true roots, and every 

 subsequent British student has confirmed Bowman's accurate 

 determination. 



M. Lesquereux informs me that his American experiences have 

 convinced him that Sigillaria is Lycopodiaceous. Dr. Dawson 

 has now progressed s s far in the same direction as to believe that 

 there exists a series of Sigillarian forms which link the Lepido- 

 dendra on the one hand with the Gymnospermous exogens on 

 the other. As an evolutionist I am prepared to accept the 

 possibility that such links may exist. They certainly do, so far 

 as the union of Lepidodendron with Sigillaria is concerned. I 

 have not yet seen any from the higher part of the chain that are 

 absolutely satisfactory to me, but Dr. Dawson thinks that he has 

 found such. I may add that Schimper and the younger German 



1 " Memoir "xi. Plates xlvii. — lii. a Idem. PI. xlix. Fig. 8. 



3 "Traite* de Botanique," p. 1304. 



school have always associated Sigillaria with t'ae Lycopodiacea . 

 But there are yet other points under discussion connected with 

 these fossil Lycopods. 



M. Renault affirms that some forms of Halonia are subter- 

 ranean rhizomes, and the late Mr. Binney believed that Halonia 

 were the roots of Lepidodendron. I am not acquainted with a 

 solitary fact justifying either of these suppositions, and unhesi- 

 tatingly reject them. We have the clearest evidence that some 

 Halonia at least are true terminal, and, as I believe, strobilus- 

 bearing, branches of various Lepidndemlroid plants, and I see 

 no reason whatever for separating Halonia regularis from those 

 whose fruit-bearing character is aim 1st absolutely determined. 

 Its branches, like the others, are covered throughout their entire 

 circumference, and in the most regularly symmetrical manner, 

 with leaf-scars, a feature wholly incompatible with the idea of 

 the plant being either a root or a rhiz mie. M. Renault has 

 been partly led astray in this matter by misinterpreting a figure 

 of a specimen published by the late Mr. Binney. That specimen 

 being now in the museum of Owens College, we are able to 

 demonstrate that it has none of the features which M. Renault 

 assigns to it. 



The large round or oval distichou sly-arranged scars of Ulo- 

 dendron have long stimulated discussion as to their nature. 

 This, too, is now a well-under.stood matter. Lindley and Hutton 

 long ago suggested that they were scars whence cones had been 

 detachel, a conclusion which was subsequently sustained by Dr. 

 Dawson and Schimper, and which structural evidence led me 

 also to support. 1 The matter was set at rest by Mr. d'Arcy 

 Thompson's discovery of specimens with the strobili in situ. 

 Only a small central part of the conspicuous cicatrix character- 

 ising the genus repre^ented the area of organic union of the cone 

 to the stem. The greater part of that cicatrix has been covered 

 wiih foliage, which, ouing to the shortness of the cone-bearing 

 branch, was compressed by the base of the cone. The large 

 size of many of these biserial cicatrices on old stems has been 

 due 1 1 the c msiderable growth of the stem subsequently to the 

 fall of the cone. 



Our knowledge of the terminal branches of the large-ribbed 

 Sigillaria is still very imperfect. Palaeontologists who have 

 urged the separation of the . From the Lepidodendra 



have attached weight to the difference between the 1 mgitudinally- 

 ridged and furrowed external bark of the former plants, along 

 which ridges the leaf-scars are disposed in vertical lines, and the 

 diagonally-arranged scars of Lepidodendron. They have also 

 dwelt upon the alleged absence of branches from the Sigillarian 

 stems. I think that their mistake, so far as the branching is con- 

 cerned, has arisen from their expectation that the branches must 

 necessarily have had the same vertically-grooved appearance, and 

 longitudinal arrangement of the leaf-scars, as they observed in 

 the more aged trunks ; hence they have probably seen the branches 

 of Sigillaria without recognising them. Personally I believe this 

 to have been the case. I farther entertain the belief that the 

 transition from the %'erlical phyllotaxis, or leaf arrangement of 

 the Sigillarian leaf-scars, to the diagonal one of the Lepidodendra 

 will ultimately be found to be effected through the subgenus 

 Favularia, in many of which the diagonal arrangement becames 

 quite as conspicuous as the vertical one. This is the case even 

 in Brongniart's classic specimen of Sigillaria elegans, 1 mg the 

 only fragment of that genus known which preserved its internal 

 structure. The fact is, the shape of the leaf-scars, as well as 

 their proximity to each other, undenvent great changes as Lepi- 

 dodendroid and Sigillarian stems advanced from yauth to age. 

 Thus Presl's genus Bergeria was base 1 on forms of Lepidoden- 

 droid scars which we now find on the terminal branches of 

 unmistakable Lepidodendra? The phyllotaxis of Sigillaria, of 

 the type of S. occulala, passes by imperceptible gradations into 

 that of Favularia. In many young branches the leaves were 

 densely crowded together, but the exogenius development of the 

 interior of the stem, and its consequent growth both in length 

 and thickness, oushed these scars apart at the same time that it 

 increased their size and altered their shape. We see precisely 

 the same effects produced upan the large fruit scars of Uloden- 

 dron by the same causes. The Carboniferous Lycopods were 

 mostly arborescent, but some few dwarf forms, apparently like 

 the modern Selaginella, have been found in the Saarbriicken 

 coal-fields. Many, if not all, the arborescent forms produced 

 secondary wood, lay means of a cambium-layer, as they increased 

 in age. In the case of some of them 3 this was done in a very 

 rudimentary manner, nevertheless sufficiently so to demonstrate 



1 " Memoir " ii. p. 222. a See " Memoir " xii. PI. xxxiv. 



3 E.g. L. HarcourtU, "Memoir" ix. PI. xlix. Fig. xx. 



