The development of the isidial outgrowth has been described by Rosendahl[509] in several species of _Parmelia_. In one of them, _P. papulosa_, which has a cortical layer one cell thick, the isidium begins as a small swelling or wart on the upper surface of the thallus. At that stage the cells of the cortex have already lost their normal arrangement and show irregular division. They divide still further, as gonidia and hyphae push their way up. The full-grown isidia in this species are cylindrical or clavate, simple or branched. They are peculiar in that they bear laterally here and there minute rhizoids, a development not recorded in any other isidia. The inner tissue accords with that of the normal thallus and there is a clearly marked cortex, gonidial zone and pith. A somewhat analogous development takes place in the isidia of _Parmelia proboscidea_; in that lichen they are mostly prolonged into a dark-coloured cilium. In _Parmelia scortea_ the cortex is several cells thick, and the outermost rows are compressed and dead in the older parts of the thallus; but here also the first appearance of the isidium is in the form of a minute wart. The lower layers (4 to 6) of living cortical cells divide actively; the gonidia also share in the new growth, and the protuberance thus formed pushes off the outer dead cortex and emerges as an isidium (Fig. 85). They are always rather stouter in form than those of _P. papulosa_ and may be simple or branched. The gonidia in this case do not form a definite zone, but are scattered through the pith of the isidium. Here also should be included the coralloid branching isidia that adorn the upper surface and margins of the thallus of _Umbilicaria pustulata_. They begin as small tufts of somewhat cylindrical bodies, but they sometimes broaden out to almost leafy expansions with crisp edges. Most frequently they are situated on the bulging pustules where intercalary growth is active. Owing to their continued development on these areas, the tissue becomes slack, and the centre of the isidial tuft may fall out, leaving a hole in the thallus which becomes still more open by the tension of thalline expansion. New isidia sprout from the edges of the wound and the process may again be repeated. It has been asserted that these structures are only formed on injured parts of the thallus—something like gall-formations—but Bitter[510] has proved that the wound is first occasioned by the isidial growth weakening the thallus. [Illustration: Fig. 85. Vertical section of isidia of _Parmelia scortea_ Ach. A, early stage; B. later stage. × 60 (after Rosendahl).] B. ORIGIN AND FUNCTION OF ISIDIA