So much of the bee keeper's success depends upon the strength of his colonies, and this in turn upon the character of the queens heading these colonies, that he needs to be well informed as to what constitutes a really good queen and how to produce such, and, having this knowledge, it will be profitable to be constantly on the alert to see that all colonies are supplied with the best queens procurable. With a queen from a poor strain of bees, or an unprolific one from a good strain, a colony, even in a season of abundant honey secretion, will give little or no return, while the seasons are not frequent during which one given a fair start and having a large, prolific queen of an active honey-producing strain can not collect a fair surplus beyond its own needs. Admitting this, it will be plain to all that queen bees differ proportionately in value as much as horses or cattle, and the keeper of bees who does not know how to select and produce the best can not be called a bee-master. When bees swarm they generally leave a number of sealed queen cells in the parent colony. With blacks and Italians there are usually 6 to 10; rarely more than a dozen. Carniolans generally construct about two dozen, but under favorable conditions can be induced to build 75 to 100 good cells at a time. Fig. 62 represents a comb from a hive of Carniolans which had built at one time 70 queen cells. Cyprians usually make 30 or 40 queen cells, but may greatly exceed this number under the best conditions, while Syrians nearly always exceed it, sometimes even building as many as 200; and the writer has seen 350 cells constructed at one time by a single colony of bees in Tunis. It might be thought that where so many were constructed only a small proportion of them would produce good queens. Such is not the case, however; for in general a much larger proportion of the cells formed by these eastern races produce well developed queens. But in all hives some queen cells are undersized. This may be because they are located near the bottom or sides, where space for full development is lacking, but in many instances it arises from the fact that they are formed last, and larvæ that are really too old to make full sized, perfect queens have to be used. These smaller cells are usually smooth on the outside and show thin walls. In selecting cells only the large, slightly tapering ones, an inch or more in length and straight, should be saved. Yet good queens may frequently be obtained from crooked cells, in case the latter are large and extend well into the midrib of the comb. [Illustration: Fig. 62.--Comb showing worker brood and queen cells. (Original--from photograph.)]