Monochromatic opacities from the Opacity Project ( OP ) ( Seaton et al . )
have been augmented by hitherto missing inner-shell contributions ( Badnell & Seaton ) .
OP Rosseland-mean opacities , \kappa _ { R } , are compared with results from OPAL ( Iglesias & Rogers ) for the elements H , He , C , O , S and Fe .
The OPAL data are obtained from the website www-phys.llnl.gov/Research/OPAL/index.html .

Agreement for H is close everywhere except for the region of \log ( T ) \simeq 6 and \log ( R ) \simeq - 1 
( R = \rho / T _ { 6 } ^ { 3 } where \rho is mass-density in g cm ^ { -3 } and T _ { 6 } = 10 ^ { -6 } \times T with T in K ) .
In that region \kappa _ { R } ( OPAL ) is larger than \kappa _ { R } ( OP ) by up to 13 % .
The differences are due to different equations of state ( EOS ) .
In the region concerned OP has the H ground state undergoing dissolution , leading to a small H-neutral ionization fraction , while OPAL has larger values for that fraction .
A similar difference occurs for He at \log ( R ) \simeq - 1 and \log ( T ) \simeq 6.4 , where OP has the He ^ { + } ground-state undergoing dissolution .

The OPAL website does not provide single-element Rosseland means for elements other than H and He .
Comparisons between OP and OPAL are made for mixtures with X = 0.9 , Z = 0.1 and Z containing pure C , O or S. There are some differences : at the lower temperatures , say \log ( T ) \leq 5.5 , due to differences in atomic data , with the OP R-matrix data probably being the more accurate ; and at higher temperatures mainly due to differences in level populations resulting from the use of different EOS theories .

In the original OP work , R-matrix data for iron were supplemented by data obtained using the configuration-interaction ( CI ) code superstructure .
The experiment is made of replacing much of the original iron data with new data from the CI code autostructure .
Inclusion of intercombination lines gives an increase in \kappa _ { R } of up to 18 % .

The OPAL website does not allow for Z containing pure iron .
Comparisons are made for an iron-rich mixture , X = 0.9 , Z = 0.1 and Z containing C and Fe with C : Fe=2:1 by number fraction .
There are some differences between OP and OPAL for that case : the OP ‘ Z -bump ’ in \kappa _ { R } is shifted to slightly higher temperatures , compared to OPAL .

Overall , there is good agreement between OP and OPAL Rosseland-mean opacities for the 6-elements , but there are some differences .
Recent work ( Bahcall et al . )
has shown that helioseismology measurements give a very accurate value for the depth of the solar convection zone , R _ { CZ } , and that solar models give agreement with that value only if opacities at R _ { CZ } are about 7 % larger than OPAL values .
For the 6-element mix at R _ { CZ } we obtain \kappa _ { R } ( OP ) to be larger than \kappa _ { R } ( OPAL ) by 5 % .