Identification of Rosellinia taxa relies primarily on morphology of stromata and ascospores, less frequently on ascal apical apparatus morphology.

The subgeneric classification adopted by Petrini (1993) is based on the morphological characters of the stroma. According to Petrini (1993), the stroma, covered and/or surrounded by the subiculum, comprises the ectostroma and the entostroma, encasing the perithecium.

The shape of stromata is usually subglobose. Whether the top of the stroma is flattened or conical may be diagnostic between some species.

The ectostroma may be hard and carbonaceous, up to 75-100 µm thick in subgenus Rosellinia, associated with a persistent subiculum. It is thinner, 25-50 µm thick, in subgenus Calomastia, associated with a non persistent subiculum. In both above cases, the stromatal surface is smooth or finely roughened. In contrast, the ectostroma in subgenus Corrugata is thin, rarely over 25 thick, and deeply folded in lower part. This last feature is conspicuous on dried material, but easily overlooked in the field on fresh material. Members of subgenus Corrugata are likewise associated with a persistent subiculum.

The entostroma features a whitish to brown tissue enclosing the perithecium, progressively reducing and usually inconspicuous in mature stromata. Its characters are not diagnostic, neither are those of the perithecia which are as a rule papery and fragile.

The subiculum in Rosellinia, that Læssøe & Spooner (1994) name "a false subiculum" on account of its situation around but not beneath the stromata, may be coarsely wolly or can bear synnemata in subgenus Rosellinia. Its is usually felty and persistent in subgenus Corrugata, felty and evanescent in subgenus Calomastia. The colour of the subiculum may vary from light to dark purplish brow, rarely whitish, but this character is rarely diagnostic. The persistence of the subiculum is sometimes difficult to evaluate in case of old stromata of subgenus Rosellinia. Remnants of subiculum are usually present around the stromatal base and visible when carefully checked.

Ascospore morphology is the more discriminating criterion at the specific level. Ascospore size, as reported by Petrini (1993) from statistical studies, usually is a very reliable character for separating related species. Their colour, from light to dark brown, is rarely a distinctive feature. Their shape can be diagnostic in some species, narrowly fusoid to broadly ellipsoid, but the morphology of the germ slit is more informative. The germ slit may be on the more convex side but most frequently on the flattened side. It may be short (1/2 spore length or less) and straight, short and sigmoid, less than spore length, as long as spore length, straight to oblique.

Ascospores of most of Rosellinia species bear cellular appendages or slimy sheaths.

Cellular appendages may be pointed and conspicuous but are often short, blunt and inconspicuous. Their presence is frequently highly diagnostic and a major difficulty in identifying Rosellinia taxa is to assess whether they are present or not. They are refractive structures, and probably best seen in water, but in most cases finding them is unrewarding. Unfortunately, we failed to illustrate these appendages with photographs taken from a bright field microscope.

When a slimy sheath is present at the ends of ascospores as slimy caps or surrounds one side or the whole spore, it can be seen in water, but is more conspicuous when ascospores are observed in diluted Indian ink (ink:1, water:1), or heated in Phloxine or Toluidine blue and then observed in water. Slimy sheaths are visible on freshly discharged ascospores, but are difficult to observe on ascospores still in the ascus, and soon disappear in aged material.

Asci usually have stipes shorter than the spore-bearing part, but they are fragile and rarely found in good condition. Ascal apical apparati usually are higher than broad and from rectangular to urn-shaped. Their dimensions are diagnostic in only a few species for which better discriminating characters can be found.

Anamorphs in nature or obtained from cultures lack characters that are distinctive enough to be used for identification, or better than morphological characters (Petrini, 1993).