10x100m of shoreline have been sampled to determine dominant species. Within each 100m random point sampling was undertaken along the shoreline. At each observation point the dominant species was recorded (S1, S2, S3 or S4). The data we have is the total number of observations of each species at each shoreline location. The question is - is one species more frequently observed than the other?

For data - see attached.

The data is categorical, has lots of zeros and sample size is uneven.

I've played around with ANOVA & Kruskal-Wallis (using transformed mean proportion data), chi-square goodness of fit (too many zeros). If somebody could point me in the right direction here that would be much appreciated.

Thanks

John