The in vitro antibacterial activity of Ulmo 90 and manuka honey was evaluated and compared. Data obtained from the agar diffusion and spectrometric assays has demonstrated, for the first time, that Ulmo 90 honey exhibits a stronger peroxide attributable antimicrobial effect against five out of seven bacterial isolates tested compared with manuka honey. Using the agar diffusion method, on average, Ulmo honey displayed larger zones of inhibition against all MRSA strains. However, in some cases, large standard deviations were observed (Table 1). This may be accounted for by the method used to inoculate the bacteria on to the surface of the agar. Although this method has been used in previous studies , a more precise method may be to seed the agar with the test organism as described by Allen et al. (1991) .
A lower MIC was observed for Ulmo 90 honey (3.1% - 6.3% v/v) in comparison to manuka (12.5% v/v) for all five MRSA strains. Although this difference, which is one dilution, may not be significant. A previous report , showed that the MBC from Medihoney (contains manuka honey) against MRSA was 3% while ours was 3.1% v/v for Ulmo 90 for 3 of the 5 strains and 6.3% v/v for the other two. That previous report  proposed that there are differences in the susceptibility of strains of the same species, which we have confirmed for MRSA isolates. The MIC values for manuka honey may seem high (12.5%), especially when compared to Patton et al. (2006) , where the same spectrophotometric assay was used. That study used a less potent manuka honey (UMF 18+) with a resulting MIC of 6.25% v/v. However, the differences observed between that study and the current study may explain this anomaly, e.g. a different strain of S. aureus was used in that study.
The removal of hydrogen peroxide activity from Ulmo 90 was shown to have reduced its antimicrobial activity. A 25% v/v solution of the Ulmo 90 had no detectable antibacterial activity when tested in the presence of catalase, where previously a 3.1% v/v solution of Ulmo honey was both the MIC and MBC for MRSA strain 0791. This would suggest that bacterial inhibition in the previous experiments was mainly due to hydrogen peroxide generation. Although some activity was observed in Ulmo 90 at 50% v/v concentration, the same activity was seen in the laboratory synthesised honey, which may indicate that activity at this concentration may be due to other factors such as osmotic pressure or high sugar content. In contrast, while the MIC and MBC was affected, a 25% v/v solution of manuka displayed antibacterial activity in the presence of catalase i.e. this was the dilution at which both MIC and MBC was observed on the removal of peroxide activity. This finding was expected for manuka as it has been previously shown that its antibacterial activity is attributed to non-peroxide components such as MGO . As catalase is present in body tissues, this may have an effect on the in vivo activity of hydrogen peroxide-dependent honeys. However the extent of this effect is not known.
Similar to other studies, this paper presents the findings of in vitro antibacterial activity of a honey against planktonic bacteria and therefore results cannot be extrapolated to the chronic wound environment. The chronic wound harbours up to four different wound pathogens  and indeed the presence of bacterial wound biofilms compound the difficulties in understanding and managing such an environment . Within the biofilm, the characteristics of the bacteria change, so that biofilm-embedded bacteria are up to 1000 times more resistant to antibiotics than the 'planktonic' bacteria that are used to test antibiotic sensitivity . The antibacterial nature of honey is dependent on various factors working either singularly or synergistically, the most salient of which are; hydrogen peroxide (produced by the glucose oxidase added to honey by bees), phenolic compounds, wound pH, pH of honey; osmotic pressure exerted by the honey, cleansing of the wound bed by the honey, level of exudate and the frequency of application. The degree to which any one of these contribute to in vivo antimicrobial efficacy has yet to be determined. However, a recent study examining the antimicrobial properties of honey in vitro found that hydrogen peroxide, MGO and an antimicrobial peptide, bee defensin-1, were distinct mechanisms involved in the bactericidal activity of honey . In addition to its antimicrobial properties, the effects of honey on host cells may also play an important role in wound healing [17, 18]. Therefore to focus solely on peroxide in honey limits our understanding of how honey may contribute to managing the bacterial wound bioburden.