Does star formation proceed in the same way in large spirals such as the Milky Way and in smaller chemically younger galaxies? Earlier work suggests a more rapid transformation of H$_2$ into stars in these objects but (1) a doubt remains about the va
lidity of the H$_2$ mass estimates and (2) there is currently no explanation for why star formation should be more efficient. M~33, a local group spiral with a mass $sim 10$% and a metallicity half that of the Galaxy, represents a first step towards the metal poor Dwarf Galaxies. We have searched for molecular clouds in the outer disk of M~33 and present here a set of detections of both $^{12}$CO and $^{13}$CO, including the only detections (for both lines) beyond the R$_{25}$ radius in a subsolar metallicity galaxy. The spatial resolution enables mass estimates for the clouds and thus a measure of the $N({rm H}_2) / I_{rm CO}$ ratio, which in turn enables a more reliable calculation of the H$_2$ mass. Our estimate for the outer disk of M~33 is $N({rm H}_2) / I_{rm CO(1-0)} sim 5 times 10^{20} ,{rm cm^{-2}/(K{rm km s^{-1}})}$ with an estimated uncertainty of a factor $le 2$. While the $^{12/13}$CO line ratios do not provide a reliable measure of $N({rm H}_2) / I_{rm CO}$, the values we find are slightly greater than Galactic and corroborate a somewhat higher $N({rm H}_2) / I_{rm CO}$ value. Comparing the CO observations with other tracers of the interstellar medium, no reliable means of predicting where CO would be detected was identified. In particular, CO detections were often not directly on local HI or FIR or H$alpha$ peaks, although generally in regions with FIR emission and high HI column density. The results presented here provide support for the quicker transformation of H$_2$ into stars in M~33 than in large local universe spirals.
