Comprehensive studies of the bud morphology are of great significance for the willow systematics. The bud shape diversity was noticed a long time ago and used for identification of willows in the wintertime by H. Shirasawa (1895), H. Nilsson (1908), and T. Resvoll (1909). However, these authors dealt with a comparatively small number of species and did not go much into detail of bud morphology. Other authors concerned with the willow buds (Schneider 1903; Wolf 1908) had still more cursory approaches. Therefore, the bud morphology so far has not received an appropriate treatment.

Both position of buds on the shoot (either accumbent or deviating off the shoot) and especially bud shape are usually very constant characters. One should consider the overall outline of a bud, looking onto its back (abaxial) side; the apex shape, looking from the back (acute, or obtuse, or rounded) and from the side (straight, or bent to the shoot, or recurved off the shoot; either tapering into a beak or not); the extent of the adaxial surface flattening and distinctiveness of the lateral carinas (which correspond to carinas of two prophylls forming the bud scale). Buds should be measured in three dimensions: the length, breadth, and thickness.

Appearance of bud scales is also of importance: their margins on the adaxial side may be either connate or distinct. Scales with distinct margins occur only in some of the most primitive sections. The anatomy of a young bud scale is similar to that of a leaf. The only difference is that the epidermis of the scale, especially its outer epidermis, is much more cuticularized. Loose, green mesophyll is in-between the two epidermal layers; vessels are inside the mesophyll. By the fall, scales become more firm, and by the winter (or during early winter), in many species they die off, either entirely or partially (about /2 to /5 1 4 ). Dead scales become thinner, as their mesophylls dry up, and change color to more fulvous or even blackish, their veins usually turning dark. Scales die off mostly in floriferous buds. In the spring, dead scales fall off entirely and promptly. In some species, scales stay alive throughout the wintertime, quite the same as they were in the fall, except becoming more pigmented and cuticularized. Scales of this kind do not shed when their buds open and stay persistent for a long time at bases of new shoots, sometimes looking like little rings. Either persistent or caducous bud scales is quite a stable characteristic, which proved to be very helpful for distinguishing species in the leafless stage. However, this character is subject to geographical variability: the farther north in the Arctic and up in the mountains, the less is dying away of bud scales pronounced. Therefore, while this character proved to be diagnostic for species identification within any particular locality, it would not always work for the entire species distributional area.

Size (and frequently also shape) of buds is subject to change along the shoot. There are three major types of bud size (and shape) gradation along the shoot.

Type 1 (alba-type, Fig. 2). The shape of buds on the shoot is not changing or changing very gradually. One cannot distinguish floriferous buds from vegetative ones by their appearance; to find out which is which, one has to consider bud contents. Buds starting from the third one to seventh (counting from the top of the shoot) are the largest on the shoot. The uppermost bud (or the two uppermost ones) are somewhat smaller. Starting from number four (to six), bud sizes are gradually diminishing towards the base of the shoot. The majority of buds open in the spring. Lower and smaller buds give birth to weaker shoots; the lowermost and smallest buds do not open at all and stay latent. However, it is impossible to predict for sure, which bud would open and which would remain latent. Buds of the type 1 are characteristic of tall willows that flower late in the spring, their catkins borne on foliated shoots.

Type 2 (arctica-type, Fig. 3). The uppermost two or three (occasionally up to five-six) buds on the shoot are of almost identical size and shape: all are large (the very uppermost occasionally somewhat smaller). Further down the shoot there occurs a sudden pronounced change of bud size, and also sometimes shape, so that the rest of buds are much smaller, nearly equal to each other. Only the upper, larger buds get to open in the spring. Some of them give birth to floriferous shoots, others—to vegetative ones. As for the small buds, they remain latent unless exposed to a special treatment. This type of buds is primarily characteristic of arctic species with their catkins terminating normally foliated shoots.

Type 3 (caprea-type, Fig. 4). Floriferous buds are very different from vegetative ones in their size and often also shape. Although floriferous buds are generally located closer to the shoot top, the uppermost one or two are usually vegetative. Vegetative buds also occupy the lower part of the shoot and occur, one or two at a time, in-between groups of floriferous buds. Within the lower group of vegetative buds, their sizes gradually diminish towards the base of the shoot. It is impossible to predict, which of them would open in the following spring and which would stay latent. The bud arrangement of this type is a feature of many forest species that start to bloom early in the spring.

Although there are, of course, some intermediate cases, the type of bud size and shape gradation is very distinct in most species. It is a very important diagnostic character, typical of entire groups of related species (and even some sections).

When considering the contents of a bud (which does not change much throughout the wintertime), one should, first of all, compare the size of the catkin primordium and leaf primordia. Frequently, the shape of leaf primordia is also of importance, as well as manner of their venation (it may be parallelinervous!) and pubescence. Certainly, the largest buds are the most suitable to see these peculiar details. There is an opinion (see, for example, Toepffer 1925; Rechinger 1957) that margins of leaf primordia in the willow buds are revolute (vernatio revoluta), which is not true: the primordia margins are not revolute inside the bud. It is only in the springtime, when young leaves, which have already grown out of buds, but not yet fully expanded, often have their margins rolled. This rolling is particularly typical for the representatives of the section Vimen; it also occurs in Vetrix, Arbuscella, and occasionally in some other sections, although the character ceases to be constant there.

Bud sizes are also species-specific in certain limits. Usually, floriferous buds are larger in male specimens than in female ones, at least in species with bud gradation of the type 3 (and in S. caprea, they are even of a different color).

In species having the bud size gradation of the type 2, the leaf number on a shoot is usually limited. In S. polaris, S. herbacea, and some other species, there are only 2-5 leaves per shoot; in larger shrubs, like S. reinii, S. alatavica, or S. jenisseensis, up to 15-25. The limitation of the number of leaves in these species is attributed to the fact that their leaf primordia are fully preformed in buds during the fall, and in the spring leaves merely expand. This is, of course, one of the adaptations to a shorter growing season. However, the number of leaves on a shoot is not always absolutely determinate. Virgate shoots in S. nummularia and S. ovalifolia with a longer growing period have already been mentioned here. Also, a stimulus from outside may induce prolonged shoot growth and formation of new leaves even in those species that normally would have a limited number of leaves per shoot (like those mentioned above, particularly, S. reinii and S. alatavica).

A typical leaf arrangement in the willows is spiral (the angle of divergence is about /5 2 ). Occasionally (for example, in S. nummularia), the leaf arrangement approaches the distichous type or, more frequently, the false opposite type when leaves are arranged in pairs, yet the divergence angle of 2/5 is retained. An arrangement of this kind is known in a number of species from the sections Helix (subsection Purpureae) and Incubaceae, as well as in a Himalayan willow S. salwinensis.

A few lowermost leaves on a young shoot, those next to the prophylls (the bud scale), usually are small and underdeveloped. Quite often, they do not even turn green and are fugacious in the majority of species. The most correct name for these lowermost abortive leaves is cataphylls (cataphylla). The cataphylls are lacking in some arctic-alpine species, like S. reticulata, which are limited to have only 2-5 leaves on each shoot per season, and these are fully preformed in a bud. Taking a short growing season into consideration, we understand that the cataphylls are an unacceptable luxury for the like species. The cataphylls are as well lacking in the second set of shoots produced in the axils of the same-year leaves (sylleptic shoots).