The implementation of a strong mother tongue-based bilingual education policy for Amharic speakers and trilingual policy for most Ethiopian pupils, despite myriad challenges in this low-income country (e.g. Wagaw 1999; Smith 2008; Terefe 2010), appears to have been exceptionally successful over the 10-12-year period from 1994 to the commissioning of our study by the Ministry of Education in 2006. We now believe that the timing of our data collection coincided with a turning point in what had until that time been a successful trajectory of policy implementation. Increasing pressure towards earlier use of English medium posed a risk to the successful continuation of MTM. However, it is significant that within a decade, MTM education was being offered in 23 Ethiopian languages with language development in at least 36 languages, partially enabled by language planning activities decentralised to the level of each regional administration (cf. Chapter 1). Wagaw (1999) had warned that this could result in uneven development and further marginalisation of ethnolinguistic communities. This was especially in relation to more linguistically heterogeneous regions, like SNNPR (Southern Nations, Nationalities and Peoples Region) and Benishangul Gumuz, in contrast with more homogeneous Amhara and Oromiya regions, and this has indeed been the case. Yet, there have been remarkable successes, and progress in some languages has encouraged development in others. In SNNPR this has resulted in the development of 12 languages for use as medium of instruction (MoI) to grade 4, eight of which were used as MoI to grade 6 until a regional policy change in 2004, and all 12 continuing as a subject through the end of primary at grade 8. In fact, since our data collection, language teams have continued to develop an additional nine languages in this region for subject teaching, though not yet as languages of teaching and learning across the curriculum. Interestingly, in SNNPR it has been possible to decentralise language development a step beyond the regions to the Zonal Education Bureaus, and there are even cases of further decentralisation to the woreda (local government) level. To date, we estimate that language development activities thus far make it possible for at least 84% of children in Ethiopia to receive MTM even if this potential has not always been realised. Addis Ababa is a striking example of lost potential, where Amharic is used as MoI whether it is the L1 or L2, and whether or not students could be served with MT materials and programmes developed in other regions.