There have been many studies on this, and most are either inconclusive, or disagree with earlier established data. Similar to our oft-debated issue of what defines "burn time", there are debates over what constitutes "ignition". There was one recent study where wood was exposed to temperatures in the neighborhood of 150 - 250F for long periods, and indeed looked like it had been burned (was charred and blackened, looking like charcoal), but never spontaneously combusted or gave off flame. Did it reach its "ignition" point? The celulose had clearly started to pyrolyze, the normal definition of ignition, but no one ever observed the wood "burning".
Here's one very basic, if not overly simplistic, description:
http://virtual.vtt.fi/virtual/innofirewood/stateoftheart/database/burning/burning.html
While more about ingnition temperature of wood not already exposed to heat for long durations (years), one of the most interesting, and telling, summaries on the subject of ignition temperatures is Table 1 on p.72 of this paper:
http://www.doctorfire.com/wood_ign.pdf
Note that here we're talking about autoignition, not piloted ignition, and this table shows that there is an enormous range and disagreement in the temperature required to cause autoignition. Temperatures reported in this table have been
measured to be as low as 200C (392F) or as high as 714C (1317F). Clearly there was a large variation in species, sample size, and many other factors... but I would say the same for the construction of our various hearths and houses.