It is clear that some EGRET-identified radio sources are highly superluminal and/or have high radio core brightness temperatures, in excess of the nominal inverse Compton limit for synchrotron radiation. This is evidence that relativistic beaming is an important effect to consider for at least some gamma-ray loud sources.
It is also clear that some strong, flat-spectrum radio sources which have not been identified by EGRET are also highly superluminal and/or have high radio core brightness temperatures. In particular, these characteristics can be displayed at least as extremely as for the EGRET-identified radio sources.
A wide variety of behaviours can be noted when the two populations are examined. There exist examples of objects which have strongly beamed core radio emission, but have not been identified by EGRET and have highly linear projected jets at least from pc-scales to kpc-scales (e.g. PKS 0637-752; 
4.3.5). There also exist examples of radio sources which do not have highly beamed radio emission, but have been identified as gamma-ray sources and have linear jets (e.g. PKS 0521-365; 4.3.3). Some EGRET-identified radio sources have bent jets, others have straight jets. Likewise for the radio sources not identified by EGRET.
It would follow that a simple, one-to-one link between beamed radio emission and gamma-ray emission is not tenable. Any plausible model for gamma-ray emission from AGN must somehow allow for a large variation in the radio to gamma-ray spectral index from object to object, or within a given object over time.
Salamon & Stecker [1994] need to appeal to an argument of this sort to explain the highly beamed but gamma-ray quiet radio sources. The existence of gamma-ray sources in jets which appear to point away from our line of sight (e.g. PKS 0521-365) is a larger problem for their proposed model since they predict that all gamma-ray sources should have their jets only a few degrees from our line of sight.
The suggestion of Dondi & Ghisellini [1995] that the radio and gamma-ray beaming cones may be collimated to the same extent is supported by the existence of objects like PKS 0521-365 and explains how gamma-rays can be observed from jets at moderate angles to the line of sight. In their model, Dondi & Ghisellini [1995] do appeal to variability in the radio to gamma-ray spectral index to explain the beamed radio sources which have not been identified as gamma-ray sources. This model may, therefore, be feasible.
The suggestion of von Montigny et al. [1995b] that jet bending
may be a simple explanation for the difference between
gamma-ray loud and quiet objects is not supported by the
data presented here. However, this explanation cannot be ruled out for some sources such as PKS 0438-438 ( 4.3.2) which appear to have extreme jet bending. We could quite easily miss the gamma-ray beaming cone in this object but easily pick up the radio beaming cone as the jet bends dramatically.
Thus, in answer to the questions posed above, it is not clear which, if any, VLBI characteristics distinguish the EGRET-identified radio sources from those not identified by EGRET. The conclusion of Mollenbrock et al. [1996] that EGRET sources have higher observed brightness temperatures than sources not identified by EGRET is the only substantial evidence in support of a difference in the beaming characteristics between these two populations, based on VLBI properties.
However, investigations of the relationships between the gamma-ray and VLBI properties of AGN are compromised by a number of limitations. One limitation is the sensitivity of the EGRET instrument. Even though the number of known extragalactic gamma-ray sources has increased dramatically by virtue of the EGRET observations, only the brightest gamma-ray sources are being identified. In addition, although 40 extragalactic sources have been positively identified a comparable number of discrete EGRET detections (39) which lie away from the Galactic plane have not yet been identified.
From the VLBI point of view, the data on the EGRET identified sources is relatively scarce, apart for a few of the very well studied sources. In particular, many of the determinations of apparent speed carry large error bars as a consequence of the sparse data which are available. Many of the EGRET-identified radio sources only have rudimentary VLBI and VLA images available, which make the task of estimating misalignment angles difficult. In addition, an unavoidable difficulty with ground-based VLBI is that the technique is not suited to the accurate measurement of high brightness temperature objects, with the consequence that arguments based on identifying high brightness temperatures with relativistic beaming cannot be accurately quantified. Observations combining Earth-orbiting radio telescopes with ground arrays will alleviate this inaccuracy since with longer baselines space VLBI will be able to accurately measure high brightness temperatures [Hirosawa 1991]. The author is the principle investigator on a proposal which has been made part of the `blazar' Key Science Program of the VSOP Space VLBI mission: a Space VLBI investigation of a sample of gamma-ray loud and quiet blazars. Included in this sample are PKS 0208-512, 0438-436, 0537-441, and 0638-752.
As our knowledge of the extragalactic gamma-ray population improves and the VLBI observations of the counterpart radio sources become more comprehensive and of better quality it may be that links between the gamma-ray and VLBI properties of AGN can be revealed. The possibility remains, however, that even though VLBI is the technique that allows the most direct indication of the importance of relativistic beaming, as close as is possible to the supposed gamma-ray emitting region, it is still probing a spatial region orders of magnitude larger than the suspected origin of the gamma-ray emission. Thus the properties of the jet on scales observable by VLBI may bear no relation to the properties of the jet in the gamma-ray emitting region.