If I had a list of numbers (1,000 rows) and wanted to find those which would appear to be anomalies, given the rest of population...what methods could I use?

There's lots of interesting possibilities. But your results must depend on the criteria you want to use.

Years ago Emil Gumbel wrote a significant book on "The Statistics of Extremes" after many years of earlier research,

http://en.wikipedia.org/wiki/Extreme_value_theory which has a bearing on this sort of problem.

Sometime later Paul Samuelson (who won a Nobel prize, but not for that) considered the interesting and related question of "How Deviant Can You Be?"

http://en.wikipedia.org/wiki/Samuelson's_inequality, that also generated a significant literature.

There's lots of other background available on this sort of topic/problem. Some of it I guess could be coded into Excel if there proved to be sufficient interest.

Maybe also worth pointing out that some distributions don't even have standard distributions, so while Excel's StDev may well produce a number, it doesn't necessarily mean anything unless you have adequate confirming additional information.

And even for distributions that do have standard deviations, most of these are unlikely to be normal ... although you can of course assume whatever you like.