La pêche au temps des lacustres sur les lacs de Neuchâtel et de Morat (Suisse) : saisonnalité et stratégie du Néolithique moyen au Bronze final. Apports d'analyses sclérochronologiques d'une population actuelle de perches Perca fluviatilis L., 1758 du lac de Neuchâtel

A study of fish bones from four pile dwelling settlements in the Three Lakes Region in Switzerland casts new light on the importance of fishing in the diet between the Middle Neolithic and the Late Bronze Age (3900 to 1035 BC). Detailed analysis of bones and scales showed that fishing made a significant contribution to the subsistence of these populations. Sclerochronological analysis brought precise information on the timing of fishing activity. Of the species identified, only perch Perca fluviatilis L., 1758 was sufficiently abundant in the different assemblages to allow such analysis. The bones were exceptionally well preserved, but only vertebrae were considered suitable for a comprehensive study of the growth marks. To obtain reference material necessary to study annual changes in growth patterns, we collected 20 perch of Various size, from Lake Neuchâtel each month between November 2009 and November 2010. From transverse sections of the otoliths we established that the different growth patterns observed were seasonally correlated. Otoliths are rarely subject to remodelling or resorption; they can therefore be used as base to the non-destructive method for estimating the season of death based on vertebrae. Our method uses the anterio-ventral surface of vertebrae 4 to 10, taking a series of measurements of the form of the edge, as well as the size of the marginal opaque zone. However, this model is only applicable to two to six year-old perch from the Three Lakes Region. Furthermore, because of the considerable environmental and climatic changes in the region during the Holocene, the model can only be used for perch vertebrae from sites that were occupied during or after the Late Atlantic period. The temporal estimates obtained using the model are reliable but relatively broad, generally limited to seasonality. This lack of precision arises from the non-negligible individual variation in the expression of the different growth structures. The main difficulties in applying this method to archaeological material from the sites studied concerned the state of preservation of the outer margin of the vertebrae. However, we were able to show that perch were captured preferentially in winter and spring during the Neolithic period and throughout the year, except in summer, in the Late Bronze Age. [ABSTRACT FROM AUTHOR]