Comparison between LBTW95 and BBP96

Hi Jacqueline. This is the main thing I found when looking into this Ly-alpha question, and the difference between Lanzetta, Bowen, Tytler and Webb 1995 and Bowen, Blades and Pettini 1996. Basically, its a question of how you treat your non-detections.

LBTW applied an equivalent width cutoff of 0.3 Angstroms to their non-detections, which is approximately the level of their MOST sensitive detection. (I am only considering the Ly-a absorbing systems here. They have better limits for the other lines). BBP used all of their measurements, regardless of the upper limits. I have combined both data sets and plotted what I thought were the most important plots of both papers: equivalent width vs impact parameter, and impact parameter vs host galaxy luminousity.

Here is the equivalent width vs impact parameter plot using all the data, and again only using the non-detections with EW limits better than 0.3A. Note how all the non-detections at r<70 kpc disappear in the latter plot. Also note that there is a nice correlation between r and EW at the small impact parameters, which basically disappears at larger impact parameters. Stocke et al. (1995, ApJ, 451, 24) showed a similar plot (their fig 5), and pointed out that you would have to measure hundreds of systems at large r before the apparent lack of systems at small r became statistically significant.

Similarly, here is the impact parameter vs absolute magnitude plot using all the data, and again only using the non-detections with EW limits better than 0.3A. Note how the latter removes all the non detections of bright galaxies at r<70 kpc.

This table shows how the statistics work out for each sample individually:

Only using EW upper limits < 0.3 A (a la LBTW):
	LBTW		BBP		Both
r<70	5/5   =	1.00	0/0		0/0   = 1.00
r<160   9/14  =	0.64	4/4 = 1.00	13/14 = 0.93
r<300	10/23 =	0.43	8/8 = 1.00	18/31 = 0.58
r>300	1/1   = 1.00	1/6 = 0.16	2/8   = 0.25

Using ALL upper limits (a la BBP)
	LBTW		BBP		Both
r<70	5/10  =	0.50	0/1  = 0.00	5/11  = 0.45
r<160   9/29  =	0.31	4/11 = 0.36	13/40 = 0.33
r<300	10/28 =	0.36	8/19 = 0.42	18/47 = 0.38
r>300	1/2   = 0.50	1/11 = 0.09	2/13  = 0.15