5 Discussion

Although the number of new OH masers detected in our survey does not allow to make a firm statistical analysis of their physical properties, we can point out some interesting characteristics. In five sources the differences between the velocities of methanol and hydroxyl maser peaks are less than 3 km s^-1. Furthermore, in several cases the radial velocities of maser peaks are the same or very close to the systemic velocities of host molecular clouds traced by thermal CO and/or CS lines. This may suggest that the maser emission of both molecules comes from common regions and the excitation of masers is achieved in the clouds of high gas density. All new masers are associated with relatively weak FIR sources; the median flux density at 60 $\mu$m is 360 Jy. The OH emission of these objects is generally weak. It seems that the correlation found between the OH maser flux density and infrared flux density (Cohen et al. 1988) is held for new sources and the infrared photons are likely to be involved in the pumping scheme of OH molecule. The detections reported in this paper were missed in previous survey biased towards bright IRAS sources (Cohen et al. 1988). On the other hand, six OH sources are located at galactic latitudes higher than $0.6^{\circ}$, so that they were not observed during the surveys of the galactic plane (Turner 1979; Caswell & Haynes 1983b).

The present survey revealed that about 28% of methanol maser sources are not associated with the 1.6 GHz mainline masers. It is interesting to look for any differences in infrared properties of OH sources and non-OH sources. Figure 2 shows the distributions of the $60~\mu$m flux density F₆₀ in both sets of objects. 52 objects were included in this figure as we neglected objects with the upper limits for F₆₀. It is clear that the methanol masers without any OH emission are generally weaker $60~\mu$m objects than those with the OH emission. The occurrence of OH emission preferentially in objects with stronger infrared emission appears to be consistent with the correlation between OH and FIR flux densities mentioned above. In turn, the methanol masers without any OH emission are associated with weak FIR sources. This suggests that the infrared photons are less involved in pumping of CH₃OH masers (Walsh et al. 1997) than in pumping of OH masers. Figure 3 illustrates the ratio of OH to CH₃OH maser peak flux densities against the (60)-(25) colour, defined as log( F₆₀/F₂₅). The peak flux densities of the 6.7 GHz methanol masers were taken from Szymczak et al. (2000). Figure 3 suggests that the intensity of OH masers increases for objects with bluer (60)-(25) colour. The above observational evidence suggests that the OH mainline masers are not sustained in sources with weak FIR emission and OH masers can appear later than the 6.7 GHz methanol masers, when a massive star-forming region evolves from the red to blue infrared colours.

Figure 2: Distributions of the $60~\mu$m flux density in OH detections a) and non-detections b)

Figure 3: Ratio of the OH peak flux density to the CH3OH peak flux density versus IRAS colour. The circles designate OH detections, the crosses non-detections (upper limits). The horizontal arrows indicate the upper or lower limits for the colour

The large scale surveys of OH masers (Turner 1979; Caswell & Haynes 1983a, 1983b; Cohen et al. 1988) provided the most complete inventory of star formation regions. The present observations however, suggest that the number of sites of star formation derived from OH data can be underestimated by about 30%. Gaylard et al. (1994) based mainly on the observations of the southern hemisphere sources also concluded that there is a substantial population of the 6.7 GHz methanol masers without any OH maser emission. A sensitive survey of selected galactic areas in both methanol and hydroxyl maser lines would yield a more quantitative result.