The genus
Graphis was introduced by the French naturalist
Michel Adanson in 1763, in a brief protologue that framed these lichens in simple, field-visible terms. He characterised
Graphis as a thin, creeping, powdery crust () spreading like a film and marked by furrows that could be simple or branched, sometimes with the margins raised into low ridges; he also recorded small spherical bodies occupying the furrows. In the same entry he pointed readers to earlier illustrations by
Dillenius. Early work on the lirellate Graphidaceae produced an unstable and contested set of generic names: during the 19th century, dozens of generic names were proposed for these "script lichens", and authors disagreed on which characters should define genera. A widely adopted late-19th-century scheme was that of
Johannes Müller Argoviensis, who treated the non-
stromatic, lirellate taxa as four spore-defined genera—
Graphis (
hyaline, transversely
septate),
Graphina (hyaline, ),
Phaeographis (grey-brown, transverse) and
Phaeographina (grey-brown, muriform)—alongside several stromatic genera. That framework was challenged by
Edvard Vainio, who united the family in a single genus
Graphis but retained comparable groupings at the
rank of
subgenus, and it was later reinstated in a more traditional form by
Alexander Zahlbruckner. Despite early criticism that the approach was artificial, these spore-based genera remained in common use well into the 20th century, until later revisions began to prioritise
ascoma anatomy over
ascospore septation in defining natural genera. The genus
Graphis has been interpreted in both narrow and very broad ways. During the 19th and early 20th centuries, most classification in the lirellate Graphidaceae relied heavily on ascospore characters (especially spore colour and whether spores were only transversely septate or also muriform), which produced a set of "ascospore genera" used for many decades. That approach was increasingly regarded as artificial, because species that were otherwise very similar could be placed in different genera on ascospore septation alone. This problem was later framed in terms of "sporomorphs": look‑alike taxa separated mainly by spore type. Modern generic limits in Graphidaceae were reshaped by revisions that emphasised ascoma anatomy, especially the structure of the (the elongate, slit-like
fruiting bodies) and their (the ascoma wall). In the
circumscription followed by
Bettina Staiger (2002) and summarised by
Robert Lücking (2009),
Graphis is characterised by lirellae with well-developed, typically darkened that usually conceal the , an excipulum that is partly to completely , a
hymenium that is usually not , and hyaline ascospores that are iodine-positive (I+) and range from transversely septate to muriform. Under this concept, many species formerly treated in
Graphina are included within
Graphis, while other taxa once placed in
Graphis in a broad sense have been reassigned to genera such as
Acanthothecis,
Carbacanthographis,
Diorygma,
Dyplolabia,
Fissurina,
Glyphis and
Hemithecium. To evaluate which
phenotype characters help delimit species and species groups, Lücking assessed 313 accepted species using a large set of
morphological, anatomical and chemical characters analysed with numerical ordination and phenotype-based
cladistic methods. A central finding was that the most informative character complex for recognising coherent groupings within
Graphis is lirella morphology, especially how the lirellae emerge from the thallus and what kind of they retain. By contrast, several traits that have often been treated as primary "key characters" (for example labial striation, the degree of excipulum carbonisation, and patterns in chemistry) can vary widely among otherwise similar species and may shift repeatedly across the genus, so they are most reliable when interpreted alongside the overall lirella type rather than used in isolation. On this basis, Lücking outlined tentative species groups within
Graphis defined mainly by lirella types, including the
G. scripta group (about 90 species),
G. subserpentina group (c. 20),
G. acharii group (c. 60) and
G. striatula group (c. 60), as well as smaller assemblages such as the
G. nuda,
G. dussii,
G. marginata and
G. symplecta groups; most currently accepted species could be assigned to one of these clusters. The study also broadened the older "sporomorph" idea into a general "morph" framework for sets of otherwise similar species that differ chiefly in one character complex (for example differences in excipulum carbonisation, hymenial inspersion, ascospore features, or secondary chemistry). These groupings were explicitly presented as a practical, phenotype-based working hypothesis for identification and for targeting future
molecular sampling in a genus too large to be comprehensively sequenced at the time. ==Description==