Loess vegetations of the Carpathian Basin have been ploughed for a thousand years. Therefore, loess steppes of verges of the Transdanubia have nearly disappeared. Therefore, in Hungary, it is also important to find out the composition of loess vegetation. The aims of our study were to date the age of the sample verges and explore types of verges. Is there a correlation between the types, location, expanse and vegetation types of the verges? Which of the coenologically important species of the genus Festuca occur in the verges and patches of loess grassland? The studies were carried out in three areas. We examined the age of the verges and their natural condition. Floristic and coenological surveys were applied to analyse the diversity of the verges. Inflorescence morphological analysis of the dominant Festuca species in the vegetation was carried out by analysing 25 parameters. Our results showed that not only the age of verges was valuable and natural. Most of the loess species did not appear in old verges. Consequently, we should not rely only on the age of the verges to make hypotheses about its natural condition and vegetation. Based on the coenological data, there were two types of verges, which were the interspaced verge and the border verge. Based on the diversity assessment from the survey, data of the interspaced verge showed the highest diversity value, the diversity value was higher than 4. Among the studied taxa, Festuca pseudovina was separated clearly as a homogeneous group. The morphotaxonomic parameters were not enough for species identification of F. valesiaca and F. callieri, but they were not enough for F. rupicola neither, so the examination of the sclerenchyma is required. Overall, in contrast to the previous hypothesis, the age of the grassland fragments is not the most important factor, but rather their location, size, the number of species and the diversity are adapted to them.