Basic reporting

The manuscript “Competition from native hydrophytes reduces establishment and growth of invasive dense-flowered cordgrass (Spartina densiflora)” reads very well, and the information is generally well organized, although the absence of an independent “Discussion” section is somehow confusing (results and discussion are shown mixed in a single section). The statistical analyses are elementary but correct, although the number of replicates for each experimental groups is always very low.

Unfortunately, I can’t help but to find a strong conceptual problem. The authors make a coherent and relevant introduction about the importance of stress-gradient hypothesis on delimiting species and population distributions (and will benefit from including citations to Maestre et al. 2009 J.Ecol. and related work). This promising introduction leads to a reasonable hypothesis for the specific study system: “We hypothesized that seedlings of the stress tolerator, S. densifolia, would be out-competed by P. australis and T. domingensis following the general ecological theory that plants showing high stress level tolerance have a lower competitive capacity than those unable to grow in stressful conditions, but with higher growth rates.”

The problem is that I don’t think that the experimental design and ulterior work allows to properly test these hypotheses. The Authors tested growth and survival of young seedlings placed in the midst of adult competitor species. I would be very surprised that under these conditions the seedlings of any species, invasive or otherwise, could have any real chance to thrive.

If I understood this right, to test the posed hypothesis the Authors should have developed common garden one-to-one competition experiments under at least two contrasting environmental-stress levels in order to obtain survival rates and individual biomass values which could be used to calculate Relative Interaction Indexes (Armas et al. 2004 Ecology) among competing species.

In its current form, the experiment is designed to confirm the observed pattern in nature; i.e. undisturbed natural habitats are not invaded by S. densiflora, an only altered or disturbed environments are sensitive to invasion. Although I can see the difficulty and work invested in reproducing exactly field conditions in the greenhouse, the results offer little ecological insight in their current form. Of course it would be very interesting to know why undisturbed native communities are resistant to invasion by S. densiflora, but this paper is not able to answer that question, since no mechanistic explanation is provided.

Minor comments:

-Common names vs. scientific names. I would suggest to use scientific (latin) names thorough the manuscript conceptually, and specifically in the title, where the common name could be removed. Common names are confusing and meaningless to most readers, particularly since they tend to be different in different regions of the world. For instance, the authors use at least two different common names in the manuscript, which only adds more confusion to it (South American cordgrass, and dense-flowered cordgrass).

-Data analyses: I suggest using total biomass and root/shoot ratios, which could be more informative, or at least complement, the independent root and shoot values.

Experimental design

The statistical analyses are elementary but correct, although the number of replicates for each experimental groups is always very low.

Validity of the findings

In its current form the presented work just replicates field conditions and gets the same pattern found in the field, but have no chance to find a mechanistic explanation for the pattern.