Wing-length is difficult to measure reproducibly, and to compare between live birds and museum skins. Recently the length of Primary 8 was suggested as a measure of wing-length in live small passerines; smaller variation was found between observers than in wing-length. This paper examines which feather-length out of Primaries 1 and 6-9 best represents wing-length in 51 passerine species. It was found that Primary 8 represents wing-length best and forms about 75.5% of wing-length irrespective of wing-shape. A formula converting feather-length into wing-length is given and tested on other samples. A new method is presented for reproducibly measuring feather-length in museum skins with several advantages over wing-length. It is shown that, in small passerines, only very small differences in feather-length occur between individuals measured live and later as museum skins.The advantages and difficulties of feather-length measurements are discussed. Feather-length is recommended as a measure of wing-length. W ing-length is the measurement most frequently taken on live or freshly dead birds and on museum skins. It serves as a measure of size in a variety of contexts, for example in analysis of geographical variation and in the distinction of sex and age. [1][2][3][4][5][6][7] Wing-length is difficult to measure reproducibly in live birds and there are systematic differences between observers. 2' 8' 9 This results in different recommendations (i.e., to measure an unflattened, a flattened, or a flattened and straightened wing s-13). Apart from the potential hazard for a live bird to have its maximum wing-length measured by an inexperienced person, 12 measurements are dependent on the experience of the observer. 14-1e Even measurements taken by experienced persons using the same method may differ significantly between observers,2 14-18 although frequent checks between measurers can result in comparable measurements. 9' 19 Measuring wing-length in skins is even more difficult. Apart from the problems already mentioned, the manner of preparing the skin results in the wings being more or less bent 9 maximum wing-length of a wing sewn on to the body is extremely difficult to measure without damage to the skin.Comparing wing-length measurements from live or freshly dead birds with those from skins and freeze-dried specimens is not possible without correction for shrinkage in museum specimens. 7, 72,[20][21][22][23][24][25][26][27][28][29][30][31] Correction factors, however, have to be determined for each species; they cannot be estimated with accuracy from other species and are affected by the preparation technique ;9' 24, 25, 30, 31 even wings of the same species have been found to shrink differently. 9,22,24,25,27,28,31 Due to this lack of comparability, especially for investigations of geographical variation, museum collections are rarely used by field ornithologists 2 L. Jenni and R. Winkler for comparison with their wing measurements.The same problems32 arise with wing formulae and indices. [33][34][35][36][37][38...