By looking at the critical period of 1989-1993, when migration dynamics and industry changed dramatically, one can gain insight on the ethnic structure of migrations by analyzing the migration patterns of Tbilisi's population. This analysis was first given in an article by recent author and English geographer Dr. Michael Bradshaw [Gachechiladze& Bradshaw, 1994]. For this study, information was gotten from the Tbilisi City Office of State Statistics (information after mid-1993 was not available as a fire destroyed it).

The total migration balance in Tbilisi (according to the propiska system) during the last eleven years is shown in Table1 .

The gradual reduction of the positive net-migration can be traced from 1986. From 1991 net-migration turned negative. The out-migration of Tbilisi's population peaked in 1993 when the socio-economic crisis in Tbilisi peaked (as it did in the rest of Georgia). The situation in 1993 was close to catastrophic: civil wars, a dramatic fall in economic output, energy shortages, general public destitution, hyperinflation, criminal "revolution", waves of IDPs, etc. In the following years out-migration waves slowed, but in-migration did not increase accordingly. During the last 11 years, 43.2 thousand more officially de-registered from Tbilisi than registered to enter. However, this does not in fact mean that the actual population of Tbilisi decreased (see below).

The data is Table 1 pinpoints the critical period in 1990-1991 when net-migration changed from positive to negative. The data of ethnic migration during this period (more exactly from about 1989 to the first half 1993) gives us a good opportunity to consider the specific dynamics of migration. In Table 2  only the net-migration indices are given according to each of the major ethnic groups.

It is obvious that the non-Georgian population out-migrated from Tbilisi more intensively than did ethnic Georgians. The majority of those who left the city were representatives of the Slavic ethnic groups. In addition, the percentage of the Slavs among the out-migrants was steadily increasing during the period under observation. As absolute figures are not reflected in the Table 2, they are presented here:

• in 1989 up to 4 thousand Slavic people de-registered from Tbilisi (they constituted 28% of all out-migrants), in 1990 - 5.3 thousands (37%), and in 1991 - more than 9 thousands (46%).

• During this period the out-migration of ethnic Ossetians increased as well: in 1990-544, in 1991-2474.

The absolute majority of out-migrants from these ethnic groups (94% of the Slavs, 88% of the Ossetians) moved outside of Georgia, i.e. actually emigrated.

In early 1992, by issue of decree from the new government, which just supplanted the former (ousted after the military coup) those citizens who rented a flat from the state gained the right to privatize the housing where they were registered. The decree was of special concern for the urban dwellers because the majority of state-owned (including cooperative, municipality and ministry-owned) housing was concentrated in the cities, while in the rural area almost all the houses were already privately owned. The decree provided potential emigrants with the chance to legally sell their house or flat (in previouse years, this would have violated the law and ended for quite a few with the loss of their state-owned flat). Incidentally, this decree stimulated the emergence of a market of relatively cheap housing in 1992-94, and many immigrants and some IDPs managed to acquire housing in Tbilisi. A similar process has been described in other former Soviet Republics [Zaionchkovskaia, 1994, 23]. Realty speculators even bought large numbers of flats. We can assume that the possibility to sell their housing hastened the decision-making process of many potential emigrants. Probably this was one of the reasons why the emigration wave was still rather high in 1992 and 1993. The number of Slavic emigrants increased to 10 thousands in each of these years, although the political factors for emigration eased.

It must be noted that the data in Table1 and 2 concerns all migrations (external and internal). from those who entered Tbilisi (attained propiska) in 1989-1993, 51.8% were inhabitants of Georgia, while the rest (48.2%) moved to Tbilisi from other Soviet Republics (FSU). Very few people moved to Tbilisi from the "far aboad". But among the out-migrants (those who abandoned their propiska) only 23.8% moved to other urban and rural areas of Georgia. The rest (76.2%) moved outside of Georgia and became true emigrants. Thus "leaving Tbilisi" in those years predominantly meant "emigration". This was even more true of the ethnic non-Georgians among whom 91.4% emigrated out of Georgia. Even a significant number (32.7%) of ethnic Georgians who de-registered from Tbilisi intended to leave Georgia.