## Analysis of multiple dependent variables of different sample sizes

Thanks in advance for your help, I'm not a statistician, so I'm struggling to understand the best way to analyse some biological data, and I apologise, therefore.

I have read through the John MacDonald advice for biologists on correlation, regression, and analysis of covariance, and I've googled and read around these subjects, but I cannot find a test that seems to fit with our data.

The data can be described...

For each of several species (species A, B, C etc) we measured:
Continuous Variable 1 on 24 individuals
Continuous Variable 2 on 180 individuals
Continuous Variable 3 on 180 individuals
Continuous variable 4 on 9-16 individuals
Continuous Variable 5 on 9-12 individuals

and

Calculated Variable 6: Variable 6 = (Mean of Variable 4 * 3) * Mean of Variable 5

The measurements were destructive, hence the individuals are different for each variable and measurement.

We wish to explore the degree to which each of the variables is correlated to species, both separately and interactively. We can plot simple correlations of the mean of each variable against species and calculate r for the answer to the separate correlations.

Could anyone advise how we should analyse the data to best determine the effect of species on the variables in each combination? E.g are there correlations between Variables 1, 3 and 4 for each species; or between Variables 2 & 3 and 5? Our hypothesis is that some of the variables are correlated to species and that there are some interactive effects between them.

We have Minitab available. However, its analyses such as ANOVA, ANCOVA, and Regression appear to require that the number of individuals tested is the same for each variable, or nearly so, and seem to imply that the different measurements are matched to single individuals.

If anyone is prepared to look at this and offer advice I would be very grateful indeed.