Ah its not so bad. If you have to take something radioactive, thats the one element I'd take because I'd pee most of it out in about a week.

The phrase no one on HN has mentioned yet is "biological half life". There's a pretty decent wikipedia article at

http://en.wikipedia.org/wiki/Biological_half-life

Whats important is the ratio of biological half life to radiological half life. Superficially it sounds like you're WAY better off with contaminated water, with a bio half life of about a week, than plutonium around 100 years, a difference of about 4 ords of magnitude. However the half life of 3H is in the 10s of years range and non-RTG Pu is about 20K years range or about 3 ords of magnitude lower activity. So the net effect is it hangs around 4 ords of magnitude longer, but specific activity is 3 ords of magnitude lower, so given X number of atoms, Pu is only about 10 times worse for you than 3H.

All bets off for the Pu RTG isotope that is so short half life it self-melts pretty easy, or that one Pu isotope thats mostly harmless but its decay product is icky (I think 241pu to 241Am where the 241Pu if it wouldn't decay is mostly harmless but the 241Am is icky)

Well, the question is where is the tritium. If it comes in the form of heavy water, then the biological half life is indeed short. But tritium can substitute essentially any hydrogen, and so it can also come in the form of amino acids, which have a good chance to end up in the nucleus of a cell. So especially for tritium, it can essentially end up in anything where organic chemistry is involved, which is to say everywhere.

However, the general problem I have with articles about the dangers of radioactivity is, that they almost never contain enough information to actually judge how dangerous it is. It is usually easy to get a factor of 100 or more variation in the dangers of radioactivity, based on the exact contamination scenario, the precise isotope abundances etc. So even Pu is not always very bad, since PuO is almost insoluble in water, which can be used for nice party tricks [1]:

     [...] to eat as much plutonium as any prominent   
    nuclear critic will eat or drink caffeine. My offers 
    were such as to give me a risk equivalent to that faced 
    by an American soldier in World War II, according to my 
    calculations of plutonium toxicity which followed all 
    generally accepted procedures.

[1] http://www.phyast.pitt.edu/~blc/book/chapter13.html

OK you motivated me to look up the biological half life of tritiated water which is about the same as regular water, it doesn't isotopically partition very much in the body.

http://en.wikipedia.org/wiki/Tritiated_water

There's an interesting wikipedia article about total body water where you basically figure out how much water is in your body in an absolute sense by consuming some D2O (or I suppose for those who want to live on the wild side, T2O) and the results keep coming up as a half life around ten days in the body. It just doesn't bioaccumulate. I think you're thinking of the heavy metal boneseekers or radioiodine accumulating in the ... gland thats by your throat (I forget?)

You are correct that mathematically you never really excrete all the T2O. However once the level decays below a banana equivalent dose, it just doesn't matter anymore.

http://en.wikipedia.org/wiki/Banana_equivalent_dose

Plus when we're talking about transuranics we have to account for the chemical toxicity which is normally much more dangerous than the radiation itself.

Great point bringing up biological half-life by the way, especially as that helps explain why some isotopes are considered more dangerous than others for what may seem to be counterintuitive reasons.