Ecology of smelt, Osmerus eperlanus (Linnaeus), from the river Thames and the river Cree

Abstract

This thesis presents ecological data for the smelt, (Osmerus eperlanus (L)), a species which has hitherto been almost totally excluded from ichthyological investigations in Britain. Samples were obtained from the river Thames on a monthly basis between February 1981- May 1982, and from the river Cree during the spawning migrations in March 1980 and 1981 and on a seasonal basis between autumn 1981 and summer 1982.

The past and present commercial value of smelt stocks has been compared and the present distribution of the species in Britain mapped. Limitations in the legislative protection afforded to smelt have been identified and the possibilities for re-introduction by means of artificial propagation discussed.

Smelt scales were found to provide a rapid method of age determination but length was found to be a poor predictor of age. Smelt from both study sites exhibited short-cycle life histories with no fish having formed a fourth annulus. There were marked changes in sex ratio with age indicating a sex selective mortality which the literature has attributed to a higher post spawning mortality of male smelt. Growth after the fall occurred at both study sites and in the river Thames the main growth period occurred between June-December. A hypothetical birth date of 1st June was assumed. Smelt from both study sites exhibited the greatest instantaneous growth in length in the first year of life, a pattern of growth typical in temperate smelt populations. The growth rate of Cree smelt (Loo = 283 mm) was much greater than that in the Thames (Loo = 187 mm) possible reasons for which have been discussed. Female smelt were generally larger than males of an equivalent age although the differences were rarely significant. Seasonal variation in condition was followed and at both sites the peak total condition corresponded with the peak in the cycle of gonad development. The development of gonads did not appear to be at the expense of somatic condition.

The seasonal pattern of gonad development was followed for smelt from both study sites and the investment in reproduction compared between sexes and between study sites. Peak gonad development occurred earlier in male smelt than in females. Insignificant differences occurred in the mean gonadosomatic ratio of Thames and Cree males but female Cree smelt had significantly higher gonadosomatic ratios than female Thames smelt of the same age. Absolute fecundity, relative fecundity and egg diameters were studied at both sites and predictive equations for fecundity were derived from multiple regression analysis. Thames smelt were characterised by higher relative fecundity but smaller egg size than Cree smelt. Within a given study site no clear relationship between egg size and fish size was evident. 21 (2.6%) Thames smelt were identified as being synchronous hermaphrodites and these individuals were compared with other hermaphroditic smelt. Smelt from the river Cree first matured at approximately 2 years of age as reported from many other populations. In the Thames however,the 1981 year class contained mature 0-group specimens presumably as a result of improved first year growth. The population dynamics of spawning smelt were monitored from the migrations into the river Cree in 1980 and 1981. The run was characterised by marked changes in size, age and sex ratio and possible factors involved in initiating the spawning migration were discussed.

The diet of smelt was found to comprise 4 major food categories - Mysidae, Gammaridae, Crangonidae and various species of young fish,and cannibalism was present in the river Cree. The diets from both study sites were significantly correlated although mysids and gammarids dominated in the diet of Thames smelt and Crangon crangon and young fish, particularly underyearling smelt,dominated in the diet of Cree smelt. Seasonal variations in the composition of the diet were marked particularly in the case of Thames smelt where mysids and gammarids exhibited alternating peaks of dominance in the diet. Possible reasons for both seasonal variations in the diet and variations in the diet with age (size) have been discussed. A new nonsubjective index of feeding activity, here termed the volumetric index, has been devised. 0-group fish were found to feed more intensively than older fish.