HI Donal:
Here are some results from the scintillator ray tracing program.
You can see that they are somewhat counter-intuitive -- the maximum
collection is not at the end.  I will investigate.

The conditions assumed are:
(1) Bulk attenuation length of plastic 3.8 m
(2) Reflectivity (internal) of surfaces 0.995
(3) All surfaces 'metal' wrapped (e.g., aluminized mylar) with
    reflectivity 0.8.

The table gives the number of photons arriving at the end of the
light pipe (i.e., the input to the PMT), as a function of distance
of the source from one end of the scintillator.  The source position
starts at 0.1 m, and steps to the center (0.8 m) of the 1.6-m scintillator
bar.  The source is centered in the cross section of the bar.
Light pipe A is at the 1.6-m end of the bar, while light pipe B is at the
0.0-m end.

There are three counts of photons: those in the primary, or direct,
group (typically the first 3 ns), those in a 30-ns window which includes
the primary group and most of the first reflected group, and all photons.

The histogram file from which these numbers were extracted can be got
by anonymous ftp, in ftp.cebaf.gov/pub/nsu/GEn/bar14.hbook

Table -- Number of photons reaching light pipe in GEn 'bar'

Source     Light pipe A        Light pipe B
Position Primary 30ns  All   Primary 30ns  All
 (m)
 0.1     1392   1998  2110   1809   2346  2390
 0.2     1475   2056  2136   1880   2423  2471
 0.3     1522   2075  2154   1963   2481  2514
 0.4     1540   2071  2144   1992   2475  2518
 0.5     1658   2181  2239   1913   2420  2469
 0.6     1749   2270  2329   2005   2547  2589
 0.7     1694   2215  2266   1825   2408  2468
 0.8     1764   2262  2305   1706   2263  2311