A student of mine carried out some research before we thought about how we were going to analyze it, and we're a bit lost.

We are trying to figure out if four particular wetland species (of two types, native and exotic) alter the the community around them. We sampled in quadrats arrayed along transects within patches of each of the wetland species in 6 ponds (randomly selected). However, not all of our species were in all ponds, and the number of quadrats per transect and transects per pond varied based on the size and shape of the area we had to work with.

Using JMP lingo I was thinking that this would be the model:

quad[transect,pond] & random
transect[pond] & random
pond & random
Species[Species type]
Species type

But is that valid, since each transect is nested within one species?