LD50 calculate a lot of ways, but this will require 3 or more group for 8-10 mice in each other. You should to look it in toxicology books or here: Antibiotic activity of two Anabaena species against four fish pathogenic Aeromonas species (http://www.ajol.info/index.php/ajb/article/viewFile/59115/4742)

But, If you need to know the approximate value, you will need only 3-4 mouse. It will be that: at the first we should to take one mouse and make a i.p. injection by 200 mg/kg (for example) and observe 40 minutes; after that one more 200 mg/kg and observe mouse behavior (convulsions, tremor, breach of locomotion, fainting) for 1 day. Then (if mouse survived) we take second mouse and do 400 mg/kg injection and etc. When approximate value becomes known we may to start study LD50 by literature methods, without unnecessary losses.

Using 5 mice per group I get a standard deviation of 10e0.3 for influenza virus LD50 in mice (http://www.pnas.org/content/98/12/6883.long) larger groups give better accuracy. I often use small groups (ie 3) and large dilutions (100 fold) to narrow down the range and then use groups of 5 mice at dilutions (1/10) spanning the LD50. I calculate using the Karber-Spearman method which is superior to the Reed and Munch.

Earl Brown, what is difference between both?

The Reed and Muench interpolates the ld50 between the 2 dosages that bracket the 50% point and thus involves the researcher inserting themselves into the calculation which moreover one degree of freedom and reduces the statistical strength of the assay. The karber-spearmen uses all the data and calculates the LD50 from all cumulative mortality.

The following 2 manuscripts compare the karber method to others

http://www.ncbi.nlm.nih.gov/pubmed/15190695

http://www.ncbi.nlm.nih.gov/pubmed/11800465

Thanks

Yusra, where can I get the software?

You should not be doing an LD50 study. In most countries this would not be allowed. Human care of animal subjects requires euthanasia when the animal shows any signs of morbidity - survival studies like LD50 are neither required nor appropriate. The NOAEL rather than the LD50 should be what you should try to determine.

5 per group according to OECD standards for toxicology study.

Minimum number 3 at least, and preferred 5 per each group