Nectary trichomes ofAbutilon striatum secrete copious amounts of sucrose, fructose and glucose. The nectar emerges from transient pores in the cuticle overlying the trichome tip cells. Calculations of the required transmembrane fluxes, either across the tip cell plasmalemma or across the cell membrane of the whole trichome, give very high rates compared with those obtained from other situations in plants and, therefore, cast doubt on the possibility that nectar secretion inAbutilon is an eccrine process. Quantitative evaluation of the possibility of granulocrine secretion, by successive fusion of vesicles with the cell membrane, suggests that this is an even less probable mechanism of secretion. Rapid freezing followed by freeze-substitution or freeze-fracture replication reveals that an extensive “secretory reticulum” (SR) is present within the hair cells. As similar micrographs are obtained from conventional, chemical fixation it is argued that the secretory reticulum is a relatively stable endomembrane system. Freeze-fracture and freeze-substitution micrographs show that this internal membrane system is closely associated with the plasmalemma. Taken together with other structural information, as well as physiological data, it is concluded that prenectar is actively loaded into the secretory reticulum of all trichome cells. Increase in hydrostatic pressure within this compartment leads to the opening of “sphincters” which connect the cisternal space of the SR to the outside of the plasmalemma. Thus a pulse of nectar is forcibly expelled into an apoplastic compartment sealed to the outside by the impermeable cuticle and on the inside by the plasmalemma. As this apoplastic compartment is also sealed at the stalk cell, the only route for pressure release is via the transient pores which overlay the tip cell. Distension renders these patent so that, again, pulsed secretion is observed. This hypothesis overcomes the necessity for envisaging excessively high transmembrane fluxes or rates of vesicle fusion. It would imply the need for a continuing supply of prenectar to the hair cells accompanied by active loading into the SR. This loading process may well be supported by the hydrolysis of sucrose to glucose and fructose and is probably the site where ions and other low molecular weight solutes are filtered from the nectar.