It is a common belief that reducing transmission power on a sensor can reduce the energy cost on transmission and therefore reduce the overall power consumption in Wireless Sensor Networks (WSNs). This belief has motivated many studies on transmission power control (TPC). An examination of this belief is presented in this paper, questioning the conditions where TPC can benefit a network. On a single link, we discover there is a limited saving achievable on a sensor itself from reducing transmission power, contradicting previous research. In a star network, however, reducing transmission power can decrease the total energy consumption on other sensors. We conclude that TPC can save energy in a certain network not through reducing a sensor it self's cost as the common belief leads us to expect, but through reducing other sensors' physical and MAC layer receiving.