In a recent paper , McLure & Jarvis ( 2002 , astro-ph/0204473 , v.1 ) reanalyze AGN broad-line reverberation measurements presented in Kaspi et al .
( 2000 ) .
They find the broad-line region size and the AGN luminosity at 5100 Å can be related by R _ { BLR } \propto L _ { 5100 } ^ { 0.5 } .
This differs from the result of Kaspi et al. , who found R _ { BLR } \propto L _ { 5100 } ^ { 0.7 } , a departure from expectations for a constant ionization parameter .
In this note I investigate the source of the discrepancy .
McLure & Jarvis use quasar luminosities based on the single-epoch , multichannel-spectrophotometer measurements of Neugebauer et al .
( 1987 ) , obtained in 1980 .
I show that the Neugebauer et al .
fluxes are systematically higher , by a constant flux offset , compared with the multi-epoch CCD measurements of Kaspi et al. , obtained concurrently with the echo-mapping radii .
In addition , McLure & Jarvis erred when converting these fluxes to luminosities .
The two effects compound to a typical factor 2 ( 0.3 dex ) overestimate of the luminosities of the PG quasars in the sample .
Since McLure & Jarvis did adopt the Seyfert luminosities in Kaspi et al .
( 2000 ) , they obtained a slope that is flatter by 0.2 .