The height of the tropopause at tropical Pacific stations shows a marked annual variation, together with a secular variation in the annual average values. Previous authors have commented on the positive correlation between the annual average height and the sunspot number. The existence of this correlation is confirmed for the period 1952--73, and a mechanism to relate soalr activity to tropical tropopause height is proposed. The mechanism, which is a simple extension of one that will be discussed in detail elsewhere, explains the annual variation in tropopause height as a response to the annual variation in surface insolation and hence in the intensity of tropical cumulus convection and of the ascending branch of the Hadley cell. The correlation with the sunspot cycle can be explained if the solar ''constant'' undergoes a fractional increase of about 0.5% from solar minimum to solar maximum.