In the 1990s it became clear that the freshwater pearl mussel, Margaritifera margaritifera (Linnaeus) was becoming very rare throughout its world-wide range. However, although information was available on its general status in some countries (e.g. Finland; 1 Ireland 2 ), these data were not at all comparable and so no reliable assessment of its true status was possible. 3 However, Scotland was considered to be a stronghold for the mussel 3 and so Scottish Natural Heritage commissioned a full, national survey of historically known mussel rivers. They required that this should produce an accurate and replicable assessment of the numbers of mussels in each river, with a view to comparing the results both with data from other countries and with future surveys. Young was commissioned to develop the survey method 4 and subsequently (with Cosgrove, Hastie and Henniger), to carry out the survey. 5,6 Since it was impossible to count every mussel, or to survey the entire length of each river, a sampling approach was taken. This was designed to be easy and quick; to be applicable to most rivers; to provide actual numbers for comparison; and to include juvenile mussels. A lack of juveniles is the most worrying feature of most pearl mussel populations. 7 This protocol is now used routinely in surveys in Scotland by Scottish Natural Heritage. A river is chosen for survey if it has old mussel records, or if its general nature (including the presence of salmonid fish, neutral or acidic bedrock, moderate slope and a varied substrate) is consistent with mussel occurrence. Searches are conducted in those areas in each river where any historical evidence exists for mussel occurrence, and in any other potentially suitable areas (based on a moderate gradient that provides a substrate including clean, coarse sand, often stabilised by cobbles or boulders). If possible, areas from upper, middle and lower reaches are included. At each general location surveys are concentrated on areas which are most likely to harbour mussels, based on knowledge of their micro-habitat requirements. 8 This is achieved by first making a general assessment of the substrate types present, both by walking up the river bank and by wading quickly in the river. Searches are made only in bright viewing conditions in low, clear water. A glass-bottomed ‘viewing bucket’ is used, to reduce surface reflections. To ensure surveyor safety, searches are restricted to safe wading depth and surveyors always work in pairs. The water is entered at the chosen site within the selected section and searching includes all suitable mussel habitat, using a slow upstream ‘zig-zag’, with concentration on especially favourable habitat. This is to maximise the chances of finding mussels. Negative results are only reported after two man-hours search of optimal habitat in the river, followed by a search for dead shells in flood debris at the riverbank. In order to provide general information on the habitat preferences of the mussel, at every site, including negative ones, habitat features are recorded. These include precise location and date; weather and river conditions; typical river width and depth; general substrate type (based on a ‘Wentworth’ scale 9 assessed by eye); main adjacent land-use; general nature of bankside vegetation; evidence of disturbance features; a sketch map, indicating area surveyed; and a photograph. If any mussels are found then continue as follows. Choose a random start point, at least 1m from the bank (so as to avoid areas likely to have been exposed in very low flow conditions), within the general area of mussel occurrence. Lay out a 50 m � 1 m transect upstream, traversing the wade-able section of the river. Work slowly upstream along this transect, counting all immediately ‘visible’ mussels. Experience has shown that 50 m transects produce analysable numbers of mussels and that 50 m accords roughly with the typical scale of mussel aggregations. 7 At 10 m and 30 m lay out a 1 � 1 m quadrat and count all visible mussels; then remove all loose stones, weeds and mussels in each quadrat and gently disturb the underlying sediment, so as to reveal buried, ‘hidden’ mussels. (Juveniles are generally buried. 7 ) The use of two quadrats is a compromise between the need to search for buried and juvenile mussels and the time consuming nature of this process. Measure all mussels along the longest shell axis and then replace safely, using a sample of 100‐500 mussels. Since size is an approximate proxy for age, this size structure provides evidence of the success of recent recruitment. Record specifically the presence of juvenile mussels (in Scotland taken to be those below 65 mm length) and of recently recruited specimens (we use � 20 mm as a guide size). 7