In their roots, or rather in the underground structures commonly called roots, the Lepidodendrons were also remarkable. Even more symmetrically than in their above-ground branching, the base of their trunks divided; there were four main large divisions, each of which branched into two and these into two again. These structures were called _Stigmaria_, and were common to all species of _Lepidodendron_ and also the group of _Sigillaria_ (see fig. 102). On these horizontally running structures (well shown in the frontispiece) small appendages were borne all over their surface in great profusion, which were, both in their function and microscopic structure, rootlets. They left circular scars of a characteristic appearance on the big trunks, of which they were the only appendages. These scars show clearly on the fragments along the ledge to the left of the photograph. The exact morphological nature of the big axes is not known; their anatomy is not like that of roots, but is that of a stem, yet they do not bear what practically every stem, whether underground or not, has developed, namely leaves, or scales representing reduced leaves. Their nature has been commented on previously (p. 69), and we cannot discuss the point further, but must be content to consider them as a form of root-bearing stem, practically confined to the Lycopods and principally developed among the palæozoic fossils of that group. [Illustration: Fig. 101.—Transverse Section through a Rootlet of _Stigmaria_ _oc_, Outer cortex; _s_, space; _ic_, inner cortex; _w_, wood of vascular strand (wood only preserved); _px_, protoxylem group.] In microscopic structure the rootlets are extremely well known, because in their growth they have penetrated the masses of the tissues of other plants which were being petrified and have become petrified with them. The mass of decaying vegetable tissue on which the living plants of the period flourished were everywhere pierced by these intrusive rootlets, and they are found petrified inside otherwise perfect seeds, in the hearts of woody stems, in leaves and sporangia, and sometimes even inside each other! Fig. 95 shows such a root _r_ lying in the space left by the decay of the soft tissue of the inner cortex in an otherwise excellently preserved _Lepidodendron_ stem (see also fig. 101). In fig. 101 their simple structure is seen. They are often extremely irregular in shape, owing to the way they seem to have twisted and flattened themselves in order to fit into the tissues they were penetrating. No root hairs seem to have been developed in these rootlets, but otherwise their structure is that of a typical simple root, and very like the swamp-penetrating rootlets of the living Isoetes.