The missing metal problem: I. How many metals are in submm galaxies?

We use a sample of submillimetre-selected galaxies (SMGs) with molecular gas and dynamical mass measurements from the literature to put constraints on the contribution of such galaxies to the total metal budget. Compared to Lyman break galaxies (LBGs), for example, SMGs are rarer (by a factor of 10 or more), but contain much more gas and are more metal rich. We estimate that SMGs brighter than 3 mJy contain only less than 9% of the metals when we combine the observed dynamical masses (few$\times 10^{11}$ \msun), number density ($n\simeq 10^{-4}$ Mpc$^{-3}$), observed gas metallicity (1--2 x solar), and observed gas fractions (~40%) assuming a molecular to neutral hydrogen ratio of 1. Including SMGs fainter than 3 mJy, we estimate that SMGs contain at the most 15% of the metals, where our incompleteness correction is estimated from the dust mass function. Our results are strong upper limits given that high gas fractions and high overall metallicity are mutually exclusive. In summary, SMGs make a significant contribution to the metal budget (< 15%) but not sufficient to solve the ``missing metals problem''. A consequence of our results is that SMGs can only add $\approx 3.5$% to $\Omega_{\rm DLA}$, and can not be the source of a significant population of dusty DLAs.