In regions of the sky where two or more sources have overlapping point spread functions, it is often difficult to resolve the number of sources and their locations. This analysis is made more difficult by time-varying sources and nearby sources with different energy spectra. The catalog entries in such regions cannot, therefore, be considered unique solutions.
As an example, Nolan et al.(1995) list a source 0445+62 at l = 146.89
, b = 10.65
, with a 
= 4.0 and a source 0500+59 at l = 150.52
, b = 10.25
, with a 
= 4.2. The present analysis, using the same initial data, the same diffuse model, and the same software (but different interpretation of what constitutes the ``best'' positions of these and nearby sources), finds a slightly higher significance for the 0445+62 source (
= 4.7) at l = 146.76
, b = 9.64
, and a slightly lower significance (
= 3.6) for the 0500+59 source at the same coordinates found by Nolan et al. Because the latitude of the 0445+62 source found in the present analysis is less than 10
, however, the source does not appear in the catalog (
< 5.0), and because the 
for the 0500+59 source is less than 4.0, it does not appear in the catalog. Both of these sources lie close to the threshold adopted for this catalog. While there seems little doubt that sources are present in this region, the number of sources above a fixed threshold and their positions are uncertain. A unique solution to fitting the data is not always possible.
The 
values for the catalog represent the statistical significance for a single source at the given position. Because a large number of observations are included in the catalog (84 individual viewing periods; 3 summed maps), the number of trial positions is large. Modeling of the EGRET data (Mattox et al., 1995a) indicates that up to 7 of the sources in this catalog may be statistical artifacts. Sources near the threshold, especially those with indications of confused or extended emission and those appearing only in a single viewing, should be treated with some caution.
Many of the high latitude sources, especially the blazars, are strongly time-variable. The pulsars show no strong time variability (see also Ramanamurthy et al. 1995). Some galactic sources appear to be time variable. A few may be associated with AGN seen through the galactic disk. Other galactic sources may be pulsars (either unpulsed emission from radio pulsars or radio-quiet pulsars like Geminga; see Thompson et al., 1994, and Romani and Yadigaroglu, 1995, for discussions) or supernova remnants ( Sturner & Dermer, 1995; Esposito et al., 1995). Some of the steady high-latitude sources could be nearby pulsars (e.g. Mukherjee et al. 1995a). As was true for the first EGRET catalog (e.g. Grenier, 1995), the unidentified sources show few likely counterparts at other wavelengths. The identification of astrophysical objects capable of producing such high-energy nonthermal radiation remains a topic of great interest.