Identification of Viola rupestris, Teesdale Violet

[From BSBI News no. 122 (January 2013), with the Editor's permission. Slightly modified.]

Viola rupestris, Teesdale Violet, from Viola riviniana, Common Dog-violet, which often grows with or near it, should be entirely straightforward, if what follows is kept in mind:

1: *** Guides and floras (at least British ones) are very misleading in the main! Very few accurately describe the
hairs of the two species correctly, with regard to their appearance and distribution. Yet these two characters at once separate the two species! ***

2: Don’t waste time on immature or stunted plants. Initially, look for well-grown individuals.

3: It is problematic seeing characters on such a low-growing plant in situ – one needs to lie prostrate, hence blocking one’s light. Collecting a leaf and stalk from a well-grown plant to allow lens-viewing in adequate light would be acceptable. Picking the rosette might however lead to the death of the plant.

4: In its UK sites, Viola rupestris in some seasons produces very few 'open' (chasmogamous) flowers: if you rely on spotting flowers, you will certainly be walking over whole colonies of vegetative plants! Learn to look for the leaves. (However, read note 5!)

5: Identification is far easier with full-grown plants in summer/autumn. Non-opening (cleistogamous) flowers grow freely as the branches develop through the season, so look for ripening capsules (the hybrid being sterile). Even long after the seeds have been cast, the three pale brown valves remain obvious – often detectable below the plant after being shed.

So - ignoring the hybrid V. × burnatii for the moment - how do you separate Viola rupestris from V. riviniana?

First, and easily observed – the leaf-rosettes are very distinctive, and leaf-shape is an excellent indication:

Shape and posture of leaves

Viola rupestris

Viola riviniana


Ovate, apex blunt.
Heart-shaped, apex ± pointed.

Base of leaf:

Basal lobes small. Sometimes ± truncate.
Broadly cordate: basal lobes large.

Posture of MATURE leaves:

Sides of leaves raised (esp. towards base), often making a scoop or trowel shape.
Basal lobes of leaves often incurved towards tip, but may be flat.
Next, needing closer inspection with a lens: just two characters will separate all well-grown plants, and indeed most others. The type of hairs is completely different, and diagnostic: very short bristle-like hairs in rupestris (‘like stubble’), but longer slender hairs in riviniana:

Type & distribution of hairs

Viola rupestris

Viola riviniana

Leaf-BLADE (upper lamina):

Apparently hairless without close inspection.
[On basal lobes and up margins, often some inconspicuous very short hairs (note ii). More rarely, very short hairs cover whole lamina.]
A covering of hairs (note i); these are slender and may be inconspicuous, but are visibly longer (note ii) than any hairs in rupestris.
[Bend leaf over finger and view against contrasting background.]

Leaf-STALK (petiole):

Typically a dense covering of very short hairs (note ii). Effect of greyish ‘fuzz’ when dense.
Sometimes covering is less dense or patchy, or petiole is entirely smooth (even on same plant).
Typically hairless (note iii).
More rarely, a few hairs.
Note i: Coverage of hairs in riviniana may be uniform, or concentrated towards the margins. Small leaves have less area and may therefore have few hairs. Leaves on young plants may be completely glabrous.

Note ii: If you can measure them, bristles of rupestris will be mostly 0.035-0.05 mm (35-50 μm); hairs of riviniana will be mostly 0.3-0.35 mm (300-350 μm) long, i.e. approaching ten times longer.

Note iii: Some forms of riviniana are more hairy with scattered longish hairs on the petiole. Rarely, the whole plant may be densely hairy. However, such forms have not been noticed on the Crummackdale limestones.

To summarise

1) Look for well-grown individuals.

2) Check leaf
   ~ leaves ovate, apex blunt, with small basal lobes or even ± truncate =
   ~ leaves heart-shaped: apex pointed, base cordate =

3) Check
upperside of leaf:
   ~ glabrous, or perhaps with minute short hairs =
   ~ a covering of distinct longer hairs =

4) Check leaf
   ~ a ‘stubble’ of very short bristles =
rupestris (but may be completely glabrous)
   ~ typically hairless =
riviniana (any hairs present will be longer than in rupestris).

Other points

Mature leaves in V. rupestris are often concave side-to-side, as stated above, but convex lengthways.

Capsules in
V. rupestris are almost globular with the apex blunt or rounded; in V. riviniana they are more elongated, and with the apex pointed to some degree. Although several published sources state that capsules in rupestris are pubescent, they are often glabrous in Ingleborough plants, or soon become glabrous, even in plants otherwise pubescent. Teesdale rupestris has the capsule more consistently pubescent (M.E. Bradshaw, pers. comm.).

V. rupestris can develop dense ‘circular’ rosettes of many leaves with short stalks; V. riviniana has few leaves in a laxer rosette. This contributes to a distinctively different appearance in many individuals.


The large and extensive populations of Viola rupestris on Ingleborough occur scattered over the two huge whaleback limestone fells of Moughton and Norber which enclose Crummackdale.

Its habitats in this area are mostly of three types:
i) calcareous drift, where terracettes caused by downhill soil-creep and livestock trampling have sparse grass cover and open patches of bare soil which allow seedling establishment;
ii) open soil in loose screes of pea-gravel and small stones, and in open turf amongst larger stones;
iii) in small-scale ‘clitter’ (broken gravel) on exposed tops;
iv) much more rarely, in cracks in exposures of limestone bedrock, especially where these form low, broken mounds.

Optimum habitats here are very largely dominated by
Sesleria caerulea, Festuca ovina and Carex flacca with calciphile mosses as a ground layer (NVC = CG9). Common associates are Thymus polytrichus, Galium sterneri, Plantago lanceolata and Campanula rotundifolia, with less frequent Carlina vulgaris, Euphrasia confusa, Gentianella amarella and Minuartia verna. Mature plants can persist in closed turf, but seedlings require open soil to establish. The species can develop dense colonies in open patches of soil in scree and amongst stones.

Sesleria is not grazed hard enough, it becomes tussocky and excludes V. rupestris and most other species. This has become a major concern in many rocky areas where grazing stock does not penetrate. Recent partial surveys suggest that populations of Viola rupestris may have declined strongly in such areas since first discovery.

The hybrid Viola riviniana × rupestris = Viola × burnatii

Naming this hybrid in the field from immature or stunted plants remains unreliable and hence worthless.

The hybrid is sterile. Thus any plant with developing capsules is not the hybrid! In my experience the flowers abort at an early stage. Hybrid plants cultivated for many years invariably aborted the flowers, to leave a shrivelled brown ‘flower-plus-stalk’ at each leaf-axil on the branches. (On a single occasion I saw a capsule on a plant thought to be hybrid, which had partially developed, but was misshapen and empty.)

In the summer and autumn look for mature plants lacking capsules, and carefully expose any branches from the surrounding vegetation to see the aborted flowers. Bear in mind that flowers in the two species often fail to develop, so the absence of capsules does not prove the hybrid, but the presence of capsules proves the pure species.

Leaf-shape is generally intermediate, being blunter than in
V. riviniana and with larger basal lobes than in V. rupestris. Margaret Bradshaw (pers. comm.) describes the shape and posture of leaves of Teesdale hybrids as being closer to V. riviniana. Given the variation in each species, however, leaf-shape may be a pointer, but the hair type and distribution is more generally useful.

The critical character: hybrids can have a variable mixture of the types of hairs of the parent species (thus, long-plus-short, or various lengths between those of the parents), and variably distributed over the stems, leaf-stalks and -blades. However, this advice on hair-length might apply only to crosses involving the pubescent varieties of
V. rupestris. See also Jonsell, et al. (2000).

The hybrid can be patch-forming; such potentially long-lived plants are known in the Widdybank and Long Fell populations, but not yet in the Ingleborough and Arnside Knott populations, where only rare solitary (i.e. not patch-forming) hybrid plants have been identified thus far. If Ingleborough hybrids generally lack a means of vegetative reproduction, their likely limited longevity might be sufficient to explain the apparent rarity of hybrids in that population.


Jonsell, B. (ed.) (2006) Flora Nordica, Volume 6. Stockholm.
Jonsell, B.; Nordal, I.; Roberts, F.J. (2000) ‘
Viola rupestris and its hybrids in Britain’, Watsonia 23: 269-278.
Roberts, F.J. (1977) ‘
Viola rupestris Schmidt and Juncus alpinus Vill. in Mid-W Yorkshire’, Watsonia 11: 385-386.
Roberts, F.J., in Rich, T.C.G., & Jermy, A.C. (1998)
Plant Crib 1998, pp. 109-111. B.S.B.I.

[Return to the gallery]
[Return to the Homepage]