The observed penetration of neutral carbon deep into molecular
 clouds poses problems both for standard models of PDR regions and
 for interstellar chemistry. We here propose to determine how the
 abundances of C+ and O vary with radius in a molecular cloud, G34.3
 +0.2, for which similar CI and O information of high quality already
 exists. This determination of the relative distribution of C+, C, O
 and CO will help substantially towards distinguishing between
 alternative models which have been proposed to account for the
 carbon penetration. Such models have included one or more of the
 effects of enhanced uv penetration into clouds due to clumping,
 generation of internal uv photons by cosmic rays exciting molecular
 hydrogen or by decay of dark matter neutrinos, oxygen depletion,
 a C/O ratio >1, turbulent mixing and gravitational contraction.
     Early observational studies of CI and CII indicated the
 widespread distribution of both species: it was established,for
 example, that the CI/CO ratio was high into depths where the
 optical extinction was 10 mag or greater, while standard PDR models
 predict that it should become very small for Av>4. It is difficult,
 however, to establish true, rather than 'column averaged' values for
 abundances deep in clouds. Recently our study of the molecular
 cloud G34.3+0.2 suggested CI/CO declined from 0.4 near the perimeter
 to 0.04 as far in as Av=55 (Little et al 1994). Our analysis
 was aided by the spherically symmetric structure of G34.3+0.2
 and the conclusion was justified by very careful radiative transfer
 modelling of the angular structure and lineshapes of CI and several
 CO transitions.
    We propose to build on this effort by making observations across
 G34.3+0.2 in the CII 158mu, OI 63 and 146 mu lines to extend our
 determinations to the radial variation of C+ and O.