Some routine haematological methods for examining fish blood are described including haemoglobin estimation, haematocrit, erythrocyte counts, erythrocyte sedimentation rate, total and differential leukocyte counts, and cytochemical staining. Descriptions of stained blood cells are given as are the ranges and mean values for the above tests on brown trout Sulmo trutta (L.). These methods are suggested as a possible means of assessing fish health but there is a need for establishing values in health, disease and various stress conditions before their value in diagnosis can be evaluated.

A review is made of some selected literature regarding the use of haematological techniques in freshwater fish pathology. Comparisons are drawn with the techniques already widely used in human pathology for the assessment of health and for aid in diagnosis of various diseases and conditions. The need for information on normal values, and on the conditions under which these were formulated, and the factors affecting them is stressed. The use of haematological values in assessing sublethal concentrations of environmental pollutants and the possibility of chromosomal changes are considered.

The various methods by which chromosomes using lymphocytes, from Salmo trutta and Cyprinus carpio, can be collected, prepared and stained for analysis are described and evaluated.

A rise in the haematocrit value of blood from rainbow trout Salmo gairdneri (Richardson) after collection into dipotassium ethylenediamine tetra-acetate and storage is reported. Investigation using blood from brown trout Salmo truttu (L.) showed this effect can be due to a too low concentration of anticoagulant in the blood. The need for studying the effects of anticoagulants on the blood of other fish species is stressed.

A comparison is made of methods currently used in separating mammalian lymphocytes from whole blood when adapted for use with fish blood. The methods give good results but do require minor modifications. The use of Ficoll-Paque, although requiring greater expertise, did result in obtaining good recovery rates of lymphocytes in a viable conditon. The separation of cells without the addition of agents other than culture medium, while resulting in good yields of lymphocytes, was much more variable and did not result in the separation of such a pure cell line. The other methods compared were found to be unsuitable for fish blood.

Brown trout (Salmo trutta L) lymphocytes were shown to separate into two fractions on a Percoll discontinuous gradient, with 53% of the cells in the 1,07gl-' fraction. The cells from the two fractions showed equal enzyme activity when stained for acid esterase and acid phosphdtase. About 70% of the lymphocytes gave a positive enzyme reaction, which if the reaction is comparable with mammalian lymphocyte cytochemistry would indicate they were T-lymphocytes. There appears to be increasing evidence among fish for the existence of T-and B-lymphocytes, and cytochemical staining could prove a comparatively convenient method for their demonstration.

No observable differences were noted in electron microscopic studies between brown trout, Sulrno drum L., and carp, Cyprinus carpio L., lymphocytes and thrombocytes separated by the Ficoll-Paque technique. Transmission electcon microscopy showed a similar ultrastructure of fish lymphocytes to those of other reported species. Scanning electron microscopy showed that these lymphocytes studied all had a villus structure and the thrombocytes clearly showed their spindle-shaped structure.

A technique is described using separated peripheral blood lymphocytes from fish which yields cultures giving good chromosome spreads suitable for convential karyotyping and the subsequent application of banding techniques.