Most of these lists are a load of tosh. Really, there can be no answer to the question about "the most venomous snake in the world", as there are so many different ways of answering that.
Probably the best compilation of LD50 values (drop-for-drop toxicity) is by Bryan Fry at:
http://www.venomdoc.com/LD50/LD50men.htmlHowever, if you are interested in how dangerous that makes a species, LD50 figures are a complete waste of time. There are several reasons for that:
1. Different species can differ hugely in their sensitivity or resistance to a venom. Some wild snake prey items have phenomenal resistance (there is a species of spiny mouse in Israel that can survive a bite by
Echis coloratus that would put a human in ICU), others very little. A spectacular human-related example is the Australian funnel-web spider, which is very highly toxic to humans (there have been a number of fatalities from bites), and yet most other mammals laugh at it. The differences between species can be thousandfold! You cannot extrapolate from one test animal to another.
2. There is no such thing as THE LD50 for a snake species. Snake venom composition is infinitely variable at all levels (between species, between different populations of one species, between individuals within the population, and even within individuals through their lifetime). Even if you take two individuals from the same population in one day, one may have a mouse LD50 that is several times higher or lower than that of the other (and moreover, the relationship may be the other way round if you were to test it on something other than a mouse). So, in a nutshell, when you read an LD50 figure for a snake species, the only thing that tells you is the LD50 of that particular pool of venoms from those particular snakes on 20g white mice. Nothing more.
3. You need to think whether you are interested in intravenous, subcutaneous, intramuscular or intraperitoneal toxicity - the results can differ by several orders of magnitude - e.g. vipers usually look pretty lame s.c. but a lot more impressive i.v., but s.c. is probably more representative of what happens in a bite.
4. Are you interested in "most venomous drop-for-drop" or "most venomous in terms of killing potential" - if it's the latter, you need to look at drop-for-drop lethality AND the venom yield.
SO, overall, LD50s and the perpetual quest for "the most venomous" snake are really unanswerable and a waste of time. Granted, in the absence of other information, the mouse LD50 may give you some sort of very vague hint as to what is likely to be highly dangerous and what isn't, but it is nothing other than the roughest possible guesstimate. And even then... there are plenty of species at the bottom of the table in venomdocs s.c. LD50 list that you do NOT want to get bitten by -
Lachesis,
Bothrops asper and WDB to name but a few...
The one thing LD50 tests are useful for is testing antivenoms - if you want to know how good an antivenom is at neutralising a venom, then you need to get the baseline lethality of the venom for your test animals first.