For HL Tau, CRL 618 and CRL 2688, small changes were made to the expected fluxes, all within the errors of the previous values. For 16293-2422, we now have peak and integrated fluxes at both wavelengths for the first time. These are still susceptible to large scatter however and so FCFs from this source should still be used with some care. For OH231.8 and IRC 10216, new light curves at 850 microns are available, both for peak and integrated fluxes. However, the larger data sets do not support the previously published single values at 450 microns, where the scatter is too large to allow light curves to be fitted. Instead, we suggest you use the 450 micron values collated approximately monthly here, which come primarily from Uranus observations and use the nightly opacity trend fits, and so generally have lower errors.

More generally, only two fluxes at each wavelength are now supplied: the peak flux and the integrated flux in a 40" aperture. We no longer distinguish between pointing peak fluxes, general map peak fluxes and photometry fluxes because the data do not support a significant difference for any of our sources.

In an individual observation, there is no way to reconstruct the correct signal from either bolometer. Hence observations of extended objects where real flux falls on either bolometer may be badly affected. Our recommendation then is that you use the SURF command CHANGE_QUALITY to switch off the emission from both bolometers for such observations (or portions of observations, if source rotation moves flux away from these two bolometers).

In the case of point sources, the signal received by both bolometers should usually be noise only. Hence these data are probably unaffected. However, in observations with chop throws around 30" or 120", one of the two bolometers may be in the "off" position. In busy fields therefore, any flux on the other bolometer may affect the chop signal seen at the off position bolometer. For faint sources this may disrupt the value calculated for C14, the central bolometer - you should compare the value in your two "off" positions, either visually through GAIA or other packages, or using the ALLBOLS=TRUE parameter option in SCUPHOT for photometry observations.

Finally, regardless of data type, we recommend that neither bolometer is used for sky-noise removal. The two bolometers can either be switched off in CHANGE_QUALITY, or you can specify their avoidance in REMSKY.

In summary:

• This affects all data from at least May 1997.
• The two bolometers (A7 and A16, or numbers 7 and 16, on the short wavelength array) have correlated flux in a constant ratio, regardless of how much real flux each should be receiving.
• For observations of extended sources, the signal from both bolometers should usually be switched off using the SURF command CHANGE_QUALITY.
• For point sources, this is likely to be unimportant. Observations with chop throws around 30" or 120" in busy fields may be affected - you will need to check your data individually.
• Neither bolometer should be used in sky-noise removal.