Referee 3's review:

This paper uses stable isotope carbon and nitrogen isotope data to study the relationship between migration, nesting, and feeding in leatherback turtles, and in so doing, demonstrates an important application of stable isotope approaches to endangered species conservation ecology. The study is timely, the manuscript is interesting and easy to read, and the conclusions should appeal to specialists and non-specialists alike. This study complements a recent study by Reich et al. (2007; Biol. Lett. 3: 712-714; PDF attached with this review). Reich and colleagues used stable isotopes to study 'cryptic' life stages in green turtles. Together, these two studies show how stable isotope approaches can (and will continue to) contribute to questions about sea turtle biology, as well as to more general ecological hypothesis, and, in some instances such as the current paper, unique insights into the ecology of endangered species.

Authors,
The main conclusion you reach is that the population you are studying (French Guiana leatherbacks) segregates into two main feeding units based on remigration intervals of 2- and 3-years, respectively. The approach using stable isotope data is novel, and the results do appear to support this conclusion. However, the only evidence provided for this assertion is in Fig. 2, where means and SE are plotted for the 2 groups in isotope space, overlayed against geographical space. A more detailed presentation of results is necessary here, possibly in the form of a table giving n, means, SE etc. for the various groups studied (and should include data for blood and yolk). You should also present data for jellyfish in table/figure format (as well as other food sources they may have), as these data are also important for conclusions about non-feeding periods. At any rate, including such baseline data always improves the robustness of isotope-driven interpretations. Further, in Fig. 2, the x- and y-axis units and labels should be moved to the bottom and left of the graphs, respectively. Finally, more detailed presentation of results of the GLMs should be included to further substantiate your claims. At present, it is very difficult to infer how analyses were conducted, and the implications of these results.

Another important conclusion is the finding that blood and egg yolk isotope data are strongly correlated, suggesting that researchers should be able to use egg yolk (as a less invasive sampling material) to study the feeding habits of adult females. This is a very important point for a variety of reasons, and should be more strongly emphasized in the abstract (i.e. add at least one more sentence on the implications of this finding).

Lastly, you should refer to Reich et al. (2007) (see above) in the final section of your Discussion. Your paper and theirs are quite complementary, and together portray the immense scope for future stable isotope applications in conservation biology.

Details:
Line 37: change 'were' to 'are'.
Line 38: insert 'including adults when not available' at the end of the sentence.
Lines 108-119: these aims/objectives/predictions are well-thought and well-written out. But, is this really the 'first analysis'? Later on you refer to a study by Wallace et al. (1994) - I don't have that paper, but it seems there might be leatherback isotope data in there?
Lines 143-162: throughout Results section, the statistical test need only be listed on first mention per section, e.g. delete "repeated measures GLM" from lines 145 and 147.
Lines 157-162: refer to Fig. 2 in this paragraph.

Discussion general:
Again this is personal taste, but I would prefer if the main finding(s) and speculations are provided at the front of each sub-section, followed by discussion in the context of the literature (at present, the context is discussed, and each sub-section is concluded by the main result of the current study).

References:
Provide authors names for reference 16.
Check format, especially journal titles, e.g. Journal Of X And Y should be written as Journal of X and Y.

**********
N.B. These are the comments made by the referee when reviewing an earlier version of this paper. Prior to publication the manuscript has been revised in light of these comments and to address other editorial requirements.