The Latin roots for the last three months of the year (the 10th, 11th and 12th months) clearly indicate 8th, 9th, 10th. What gives?
It’s my understanding that two months were added to the original calendar. Specifically, July and August (named after two Roman diplomats - no prizes for guessing who).
The Romans started the year with March. January and February were ignored, and not named for quite some time. I agree with this system, particularly February. I hate that month!
As Annie-Xmas said, the Roman calendar started with March an was based on agriculture. The first months were Martius, Aprilis, Maius, and Junius. The next six were numbered: Quintilis, Sextilis, Septembris, Octobris, Novembris, and Decembris. There was an unnamed period between Decembris and Martius when nothing much happened in terms of agriculture. Eventually, two months, Januarius and Februarius were added, to bring the number up to twelve. The Romans also had an additional period, called Intercalius or Mercedinus, to make up the difference between the number of days in their calendar and the number of days in the solar year.
Julius Caesar eliminated Intercalius and changed the number of days in the twelve months to have the calendar match the solar year of 365. Quintilis was later renamed Julius to honor him, and Sextilis was later renamed Augustus to honor Augustus Caesar (Octavian). One of Augustus’s big contribution to the Roman calendar was changing the number of days in many of the months. What triggered this was his desire for his month of Augustus to have the same number of days as Julius. Nothing like a little ego.
Much of this information is from Bill Hollon’s calendar site. You can go directly to the Roman calendar page.
Many places chose to keep the beginning of the year in March for many centuries. England, for instance, continued to begin the year in March up until 1752, when they adopted beginning the year in January and Gregorian calendar reform at the same time. This is why, for instance, you will sometimes see George Washington’s date of birth listed as something like “Feb 22, 1732 (Feb 11, 1731 old style)”. 11 days different than the calendar actually in use when he was born because of the Gregorian change, and a different year because it’s a date in February, and the older calendar began years in March.
A huge amount of calendrical background can be obtained from Peter Meyer’s site as well:
Damnit, I wish I had the ability to retract a post. Thanks for fixing my month-stupidity. 
According to this site we do not have a clear idea of exactly when the first month moved from March to January (although it preceded Julius Caesar).
The site has a number of other good comments on the Roman Republican calendar and the Julian reform, as well.
I was born in February!! 
::sniff:: :mad: