QUASAR HOST GALAXY DETECTION USING THE CROSS-CORRELATION TECHNIQUE

We show, using simulated data, that the two-component cross-correlatio n technique is a more discriminating method of determining the morphol ogy of the host galaxy of a quasar than the chi2 fitting of model inte nsity distributions to the quasar's radial intensity distribution whic h previous authors have used. We also show that even if the signal-to- noise is not high enough to enable galaxy morphology to be determined, a very accurate measure of the contribution of the central point sour ce to the image can still be made. This enables much better measuremen ts of m(qso) and m(gal) than is possible using the methods used by pre vious authors. We apply this method to an i-band image of PKS 2349 - 0 14 and show that the use of those methods used by previous authors wou ld have led to significant errors in the measurement of m(gal) and m(q so) for this object compared to those values obtained using the cross correlation technique. Such errors could seriously affect the statisti cal relationships which have been derived between m(qso) and m(gal) fo r quasars and the relationships between the magnitudes of the host gal axies of radio-loud quasars and radio-quiet quasars and similar relati onships between quasar host galaxies and radio galaxies. We suggest th at a re-analysis of the data using the cross correlation method may af fect the conclusions reached and should be undertaken. We also suggest that applying the technique to high S/N data may enable a determinati on of host galaxy morphology at a high confidence level.