The use of food as a carrier for probiotic organisms is of considerable interest to food manufacturers due to the claimed health-associated benefits of probiotics. However, maintaining high numbers of viable probiotics in fermented milks is not easy, and a large quantity of probiotic cultures is needed to compensate for the likely losses of probiotics during the shelf life 
. Procedures for enumeration of lactobacilli have not been properly defined. Such a situation causes difficulties in quality control of the probiotic products containing lactobacillus
species using the conventional enumeration technique. The suitability of various media to selectively enumerate lactobacilli has been examined in different studies. Although there are several elective/selective media for isolation of lactobacilli, the levels of recovery of the lactobacilli are discordant with each other.
Oberg et al. (2011) reported that while MRS-Sorbitol is a medium designed for Lb. acidophilus
in which sorbitol is the sole sugar, Lb. casei
can also grow on the medium, although only at elevated incubation temperature (42 °C). At this temperature, the MRS-Sorbitol medium gave higher bacterial counts compared to the Lb. casei
specific medium (Lactobacillus casei
agar), indicating that it could be used to obtain the total LAB count at different temperature 
. However, in our study, colonies of target strains were recovered at 37 °C on MRS-Sorbitol agar. Due to the high recovery, no other recovery temperatures were employed.
MRS-Sorbitol demonstrated higher viable counts than MRS-Clindamycin, suggesting that MRS-Sorbitol might allow the growth of additional LAB. Shah (2000) stated that MRS-Sorbitol agar could not be used for selective enumeration of Lb. casei and Lb. acidophilus in products containing both bacteria.
This study also reports that MRS-IM Maltose is not an ideal choice for selective enumeration of lactobacilli since the recovery was low compared with other MRS variants.
MRS-Clindamycin has been proposed for enumeration of lactobacilli in different studies 
. Furthermore, the International Organization for Standardization (ISO) (2006) recommended MRS-Clindamycin agar for the enumeration of Lb. acidophilus
in dairy products in the presence of other probiotics including other lactobacilli, streptococci, and bifidobacteria 
. Simplicity of medium preparation and availability of the antibiotic supplement led to its consideration as the preferred medium compared to the other selective media. Moreover, for Lb. casei
to grow on MRS-Sorbitol, the incubation temperature should be raised to 42 °C, therefore it is impossible to have differentiation on one medium and at one incubation temperature 
. Hence, in our research, MRS-Clindamycin was considered as a reliable medium to selectively enumerate Lactobacillus spp.
in fermented dairy products. Having said that, the selectivity of MRS-Clindamycin may not be 100%, as S. thermophilus,
which is difficult to distinguish morphologically from Lactobacillus
spp., was also isolated and identified in sample no. 23. This was not further investigated.
Our research shows that on the purchase and the expiry dates, respectively, 86% and 61% of tested samples contained the minimum recommended therapeutic level of log10
6–7 CFU/g, concordant with the findings of the others 
. Other researchers have also reported commercially probiotic dairy products with inadequate amounts of viable probiotics 
, which in some cases may be attributable to disruption of the cold chain 
. In this study, during cold storage, the number of Lactobacillus
spp. in some samples decreased considerably. The most important contributing factors for loss of cell viability are decreasing pH during storage, presence of dissolved oxygen, and presence of preservatives in the final products 
. In this study, the pH decline between the purchase and expiry date was in some cases noticeable. It could be due to continued fermentation process by LAB even in low temperatures (post-acidification). However, no correlation was found between pH decline of samples and their probiotic counts.
The presence of dissolved oxygen might be the other important reason for drop in viability of cell count in fermented milk 
. The majority of tested products in this study were stirred yoghurts, in which air could have been incorporated when the yoghurt was mixed with the fruit compote. In addition, some of the commercial fruit products contain preservatives to control contamination and this might affect the viability of the probiotic cells 
Based on results obtained in this research, which confirmed lower counts of probiotic cultures approaching the end of shelf life, and supported by the study of Jayamanne and Adams (2006), it is recommended that probiotic fermented products need to be consumed earlier than the expiry date to ingest maximal numbers of probiotic bacteria.
Although there are no universally established standards for microbial content and health claims for probiotic products, the manufacturers should at least clearly express the genus, species, and strain of the probiotic microorganism(s) and also the minimum viable count of each probiotic strain at the end of shelf life 
. To ensure that the consumers benefit from commercial probiotic products, it is necessary to confirm the identity of the claimed organisms at species/strain level and that they are present in the product in appropriate numbers before consumption. Some of the tested products in this study presented inadequate information on the labels. Microbial investigations of probiotic products by others have indicated that the number and identity of recovered species do not always correspond to those stated on the labels of products 
Identification of probiotic species used in carrier products should be verified in support of claimed health benefits. To obtain accurate and reliable identification of the probiotic species, molecular techniques should be applied. It has been suggested that DNA profiling by PCR-based methods are the best means for identification of probiotic bacteria at strain level 
. Many misidentifications of probiotic microorganisms may be due to the use of solely phenotypic methods for taxonomic characterization 
The rep-PCR fingerprinting profile revealed relative genetic differences between the tested isolates. In this study, 85 isolates from fermented milks were grouped based on their DNA patterns by rep-PCR, and 20 isolates out of 85 were selected for identification by sequence analysis of 16S rRNA. Amplification of the 16S rRNA gene often provides a rapid and reliable tool for bacterial identification without the need for phenotypic characterization. However, 16S rRNA sequencing cannot discriminate between closely related species. Thus, sequencing of alternative genes, such as rpoA,
with more discriminatory power has been proposed 
In this research, amplification and sequencing of the rpoA gene did not provide enhanced discriminatory information for the tested isolates compared to the use of 16S rRNA gene sequences. Sequencing of other genes, such as rpoB and pheS, would enhance discriminatory potential, enabling differentiation of strains with close genetic profiles. Anyogu et al. (2014) stated that sequencing of the pheS, rpoA, and rpoB genes along with 16S rRNA gene sequencing provides a better identification of LAB and Bacillus isolate.
Even though more media have been suggested in recent years for the enumeration of probiotic lactobacilli in fermented dairy products, none seems to be suitable for all lactobacilli or at least for Lb. acidophilus/Lb. casei (which are the two most frequently used lactobacilli in the products marketed in the UK/EU), or at the same time be able to act as a differential medium for these two species. Therefore, in this study we examined and compared a limited number of media.