A suggestion to get you started: download R and have a look at the vegan package (in fact, you can access the manual from the R-CRAN without even installing R). It's a standard toolkit for community analysis.

Try with the links below may be helpful for you,

http://folk.uio.no/ohammer/past/

http://www.sams.ac.uk/peter-lamont/biodiversity-pro

You can try the modeling technique of maximum entropy distribution or Maxent. Maxent is a maximum entropy based machine learning program that estimates the probability distribution for a species’ occurrence based on environmental constraints (Phillips et al., 2006). It requires only species presence data (not  absence) and environmental variable (continuous or  categorical) layers for the study area. We used the  freely available Maxent software, version 3.1 (http://www.cs.princeton.edu/~schapire/maxent/).

A suggestion to get you started: download R and have a look at the vegan package (in fact, you can access the manual from the R-CRAN without even installing R). It's a standard toolkit for community analysis.

you could use various packages associated with the software R,  Zuur et al 2009 "Mixed effects models and extensions in ecology with R" (Springer) demonstrate some examples how to use mixed models and glm models for intertidal assemblages.

i agree with Graeme, download R with the vegan package and problem solved!

I agree with previous answers. I could add two commercial packages that are also useful in case you are not familiar with R environment:

(1) CANOCO (last version: http://www.canoco5.com/)

(2) PRIMER (http://www.primer-e.com/)

both software were widely used in intertidal/benthos community analysis 

Just to give my two cents on the mentioned software:

PRIMER: Nice surface and lots of possibilities for biodiversity research (e.g. taxonomic distinctness), but you are restricted to the measures included (e.g., I sometime need Cosine similarity as measure for resemblance which is not available); not for free.

PAST: Surface not as intuitive as PRIMER, but anyway nice and with other possibilities more related to paleontological research; Cosine similarity included; for free.

BiodiversityPro: As far as I know, for free. A bit less functions than in PRIMER, but good for most issues of community analysis.

CANOCO: Specialised software for canonical correspondence analysis and related stuff. Useful to relate species and stations to environmental factors.

R: Environment for statistical analyses of any kind. You need to learn the R language which takes some time. But then you can do (nearly) everything: Doing community analysis (package vegan); modeling species counts against fixed and mixed effects (package mgcv and others); writing scripts only once that do very complicated things and then just let them run over different data sets; perfect also for designing graphs of all kinds. And it's for free. My recommendation if you are willing to learn that language (and by the way, you find a job much easier lateron, since mastering R is often higly appreciated)

All the best

Armin

I'm fond of Ecosim.....http://www.garyentsminger.com/ecosim/

Thanks to all, for information. However, I can’t purchase some software, but tried freeware and got large numbers of analysis for community ecology. Once again thanks to all.