Redds of trout, Sulmo iruttu L., Atlantic salmon, S. sulur L., and rainbow trout, S. guirdneri Richardson, were studied at sites in north-east England, south-west Wales and southern England (Dorset). Spawning females preferred gravels of 2G30 mm diameter and water velocities at 0.6 depth of 3 15 cm s-' and < c . 2.0 body lengths s-'. After logarithmic transformation the major horizontal dimensions of redds could be related to one another by linear regressions, and redd tail length could, similarly, be related to female fish length. Egg burial depth in upland stony streams in north-east England and south-west Wales could be predicted from a linear regression of mean depth upon female length. Information is also given on gravel composition and redd structure.

A desk study of the relationship between temperature and hatching time for the eggs of five species of salmonid fishes D. T. CRISP Freshwater Biological Associaton, Teesdale Unit, England SUMMARY. Information on temperature (T'C) and time from fertilization to 50% hatch (D days) for five species of salmonid fishes has been used to assess several mathematical models relating D and T. No single equation gave the best fit to all five data sets. The power law with temperature correction (a), Iog,,|D = logt,,fl + /j logio (T-a) and the quadratic, log,oD = loginfl + fcr+ft]7^ (where a, b, bi and a are constants), each accounted for over 97% of the variance of D and were good fits to the observed data points for all five species. There was little difference between the predictions obtained from these two equations within the range of observed temperatures. Therefore, the simpler power-law model is preferred. However, there were substantial within-species differences between values of D predicted from extrapolations of the two models from 2 or 3''C down to 0*C. When more data for low temperatures become available it will be possible to make a more objective choice of model.

1, The literature contains a number of curves relating time (days) required for median hatch ( = 02) to water temperature (7",°C ) for the eggs of several salmonid fishes. There are relatively few data on the relationships between time to median eyed ( = Z),, days) or time to median swim-up (^D^, days) and temperature.2. From published data, over most of the range 0-9.5°C, approximate relationships are D,=0.5Z>2 for Atlantic salmon {Salmo salar L.) and D^=\.1D^ for eight species of Salmo and Oncorhvnchus.3. Field and hatchery tests suggest that these are useful empirical models for approximate prediction of Di and D3 from D2 for most salmonids.

Water temperatures from six north Pennine streams (approximate mean annual discharges 0.04-1.5 m' s ') and two Lake District streams (mean annual discharges c. 0.04 m' s"') were used to examine the possibility of predicting 5-day or 7-day mean water temperatures from standard meteorological/hydrological data collected at north Pennine stations. The relationship between mean air temperature (X) and mean water temperature (V) was approximately linear, except in the Pennine streams for periods when mean air temperature was below OX. The calculated regressions of mean water temperature upon mean air temperature accounted for 86-96% of the variance of water temperature even when the air and water temperatures were measured at stations some 50 km apart. A negligible improvement was obtained by the use of multiple regressions incorporating rainfall or stream discharge from a nearby metering weir. Predicted water temperatures were rather lower than observed values during the first half of the calendar year and rather higher during the second half, but these seasonal differences could not be adequately quantified tor practical application.

No abstract