Abundant research has shown that encoding meaningful gesture, such as an iconic gesture, enhances memory. This paper asked whether gesture needs to carry meaning to improve memory recall by comparing the mnemonic effect of meaningful (i.e., iconic gestures) and nonmeaningful gestures (i.e., beat gestures). Beat gestures involve simple motoric movement produced along with the rhythm of the speech (e.g., hand with open palm flips outwards). Although beat gesture does not carry any semantic meaning, it serves a meta-cognitive function by marking the parts of speech that the speaker would like to emphasise. It also activates prosodic processing of the accompanying speech. We also asked whether the mnemonic effect of both types of gestures was found in both adults and children. In both experiments adults and 4-5-year-old children watched 3 different videos, each consisted of a list of 10 words (5 words for children), in 3 conditions (words accompanied by iconic gestures, words accompanied by beat gestures, and words not accompanied by gestures) and were asked to recall the words without moving their hands. Not surprisingly, encoding iconic gesture improved memory in both adults and children--they recalled more words when encoding them with iconic gestures than when encoding words alone. Interestingly, encoding beat gestures aided recall in adults, suggesting that both meaningful and nonmeaningful gestures could strengthen their memory. More importantly, adults recalled comparable number of words when encoding them with iconic gestures and beat gestures. However, such mnemonic effect of beat gesture was not found in children, suggesting that young children might not be sensitive to the meta-cognitive aspects of beat gesture. (Contains 2 figures and 2 footnotes.)