1. A periodic adjustment of fq against fm represents the first stage of application of molecular generators for controlling and checking fq, when it was necessary to keep that frequency within certain measurement limits γq, thus satisfying the frequency measurement requirements. For the purposes of time-keeping the use of this method is not promising since it does not ensure the use of molecular clocks with an error limited to the error of molecular generators, especially if the stability of fm is further increased. 2. Of the various methods of continuous automatic adjustment or measuring of fq against fm, the APC system and integral measurements are preferable since they provide actual and analytical molecular clocks with an error which is determined over a long time period by the error of fm, thus providing the most uniform time and frequency scale. 3. The raising of the stability of fm over a long period of continuous operation can be achieved in the first place by decreasing the effect of the molecular generator parameters, by making the generators work under optimum conditions and controlling automatically the values of these parameters, i.e., the temperature of cavity circuits, voltage of the focusing device and beam pressure. 4. One of the basic requirements of a molecular as well as any other clock is its continuous prolonged operation. The simple conditions of their utilization make molecular clocks suitable for use in the time and frequency service for the purpose of keeping reference time and frequency. In this connection any further development of molecular generators should be aimed at eliminating the use of frozen ammonia and of continuously operating prevacuum pumps. 5. Further improvements in molecular clocks may lead to their becoming the basic time-keepers, working standards, of time and frequency, whose actual value will be determined from astronomical observations. These measurements are necessary because the oscillation frequency of the molecular generators is unstable and can over a long period of time lead to certain relative errors Srm of the clock. In order to measure this error against the primary standard, the ephemeris time, it is necessary to select a long period of astronomical observations so as to make Srm≅3α0, whereα0 is the relative error of the astronomical determination of this time interval. If there exist molecular clocks whose relative error over a long period does not exceed 1·10−11, the actual value of their indications would have to be determined, considering the present accuracy of astronomical measurement, over a period of several dozen years.