Mollenbrock et al. [1996] list the observed brightness temperatures, or lower limits, for 140 compact radio sources derived from 22 GHz intercontinental VLBI observations. They plot the brightness temperature distribution for their sample and highlight the EGRET-identified sources. It appears that these 21, 17 high confidence and 4 marginal, EGRET-identified radio sources prefer the high brightness temperature end of this distribution, although the brightest and highest brightness temperature source in their sample is PKS 1921-293.
Mollenbrock et al. [1996] evaluate the statistical significance of this result with a Kolmogorov-Smirnov test and find that the observed brightness temperature distribution of the EGRET-identified sources is different to that for the 119 sources in their sample not identified by EGRET at a confidence level better than 95%. This result will need to be confirmed since to measure high brightness temperatures accurately longer baselines must be used. If the cores in radio-loud AGN are, in reality, unresolved then the measured brightness temperature is independent of the frequency of observation and only sensitive to the length of the baseline. Thus, high frequency observations on Earth-based baselines are not effective for measuring high brightness temperatures, hence the need arises for the Earth-space baselines which will be provided by the upcoming VSOP and RadioAstron space VLBI missions.
Space VLBI observations have previously been made with the TDRSS satellite in conjunction with ground-based telescopes [Levy et al. 1986]. Linfield et al. [1989; 1990] report on the measurement of source frame radio core brightness temperatures from these observations.
At 2.3 GHz, Linfield et al. [1989] detected 23 out of 24 sources
observed and could derive Gaussian models and therefore brightness
temperatures for 14. Two of these sources have been identified as
gamma-ray sources and the remaining 12 have not. The
brightness temperatures of the two gamma-ray sources, 0420-014 and
1253-055, are midrange in the sample, 3.14 
10 
K and
1.59 10 K respectively. The brightness temperatures
for the remaining 12 sources ranged between 2.5 10 
K
(0723-008) and 3.80 10 K (PKS 1921-293).
0420-014 and 1253-055 were the third and seventh highest brightness
temperature objects respectively.
At 15 GHz, Linfield et al. [1990] were able to derive Gaussian
models and brightness temperatures for 9 of the same 24 sources. Four
of these are identified gamma-ray sources. Again, the brightness
temperatures for these sources, 1223-023, 1253-055, 1510-089, and
1730-130 are midrange for the sample, 1.14 10 K, 1.98
10 K, 1.55 10 K, and 8.7
10 K respectively. The remaining 5 sources are gamma-ray quiet
and have 15 GHz brightness temperatures ranging between 5.0
10 K (0823+033) and 2.40 10 K (3C 446). The
ranking of the gamma-ray sources in brightness temperature was sixth,
fourth, fifth, and seventh respectively.
The results of section 4.3, Mollenbrock et al. [1996] and
Linfield et al. [1989; 1990] are in a sense consistent. All of
these investigations have shown that the EGRET-identified radio sources
can have core brightness temperatures in excess of 10 K, the
nominal inverse Compton limit for synchrotron radiation and therefore, Doppler factors greater than unity. However, the highest
brightness temperature objects are not all EGRET-identified sources.
In particular these studies have one object in common, PKS 1921-293,
which is not an EGRET-identified source but consistently revealed to be the highest brightness temperature
object.
Despite this level of consistency, the large-sample study of Mollenbrock et al. [1996] has shown evidence that the EGRET sources, in general, have significantly higher radio core brightness temperatures than those sources not identified by EGRET, at least as measured with ground-based VLBI observations.