Key points are not available for this paper at this time.
We used factor analysis to examine the correlation structure of six multivariate blood chemistry data sets for migrating hatchery and wild juvenile chinook salmon (Oncorhynchus tshawytscha). Fish were sampled (19982002) from juvenile fish bypass systems at dams or (one data set) from fish transport barges on the Snake and Columbia rivers. Analyses were performed to determine which blood chemistry analytes covaried, to facilitate interpretation of the data sets, and to provide insight into controlling physiological mechanisms. Four underlying factors were derived from the analyses: (i) a nutritional factor composed of total protein, cholesterol, calcium, and alkaline phosphatase, (ii) a tissue damage factor composed of the enzymes alanine aminotransferase, aspartate aminotransferase, and creatine kinase, (iii) a lipid metabolism factor composed of triacylglycerol lipase and triglycerides, and (iv) a stress factor composed of cortisol, glucose, Na + , and Cl . Although causal mechanisms cannot be directly inferred from our analyses, findings of published research provide tenable causal mechanisms for the observed structure. The consistency of the correlation structure among data sets suggests that composite (latent) variables may be more reliable indicators of some physiological responses than changes in individual variables.
Wagner et al. (Thu,) studied this question.