Betula celtiberica Rothm. & Vasc.

Found in montane forests up to the tree line, above the upper limit of beech, Betula celtiberica clings to the steep mountainside, the trees bent by soil-creep and winter snow. As circumscribed in the present treatment, this tetraploid species includes only birch populations from the Cantabrian Mountains of northern Spain, though the species as first described by Rothmaler and Carvalho de Vasconcellos in 1940 encompassed birches in many of the mountain ranges of northwest Spain and eastern Portugal. It is possible that other populations of birch in isolated ranges in central Spain may belong to this species, but it seems more likely that they are relict populations of B. pendula, and perhaps also of B. pubescens (Peinado & Moreno 1989).

Morphologically, Betula celtiberica resembles diploid B. pendula rather than tetraploid B. pubescens, although it is mentioned under the latter by Max Walters in Flora Europaea (Tutin et al. 1964), probably on the basis of chromosome number. Betula celtiberica differs from European B. pendula subsp. pendula in the presence of axillary tufts of hair in the vein axils on the underside of the leaf, the greater hairiness of young shoots, and the less markedly double-toothed leaves. From B. pubescens it differs in the shape of the fruiting catkin scale, in which the lateral lobes are usually more or less recurved as in B. pendula; in the rough, warty texture of the hairless shoots and twigs; and in the rhomboid (rather than ovoid) shape of the mature leaves, which (except for the hair tufts) are more or less glabrous.

Betula celtiberica could plausibly represent a stabilised population derived from B. pubescens heavily introgressed with B. pendula, and it might be questioned whether it in fact deserves recognition as a distinct species. However, its geographical isolation, and the fact that it can be distinguished morphologically both suggest that it is worth recognising taxonomically. In Flora Iberica, Peinado and Moreno (1990) did not distinguish it from B. pubescens, but it is equally out of place in B. pendula. In fact, as can be seen from the key, it is probably most similar to another ‘species’ of stabilised hybrid origin, the east North American Betula × caerulea.

In cultivation in the British Isles the male catkins develop earlier relative to the leaves than in related species so that the tree is showy in catkin. It is one of the last birches to lose its leaves, which remain green well into November. It could therefore be expected to grow well in climates with a long, frost-free growing season where there is adequate summer rainfall. The bark is variable in colour, but in some individuals is very white, so selecting individuals with the most attractive bark might yield useful amenity trees.

A handsome tree (19 m tall in 1985) grown by Collingwood Ingram at Benenden, Kent was referred by Clarke (1988) to B. celtiberica; clonal material from this tree was propagated and may be planted elsewhere. However the tree was raised from a 1929 collection from the Gerez mountains of Portugal, so cannot belong to the species as circumscribed here.