2. With regarding-to the relatively slow rise times of the IPsCs, we think there are two-possibilities. First, more or less, the IPSCs will be unavoidably filtered by its process structure, specially their fast rise phase( see the reply 1a)using i simulation with typical values of the parameters used above. we found ifthe 10- 90%eme05 ms for IPSCs pro时动一 te /efz xke A w44 A oL-LfesABr-44 4 end of BC axon,枚 ed-down to about o,62ms(24% increase),when recorded from BC soma. However when the 10-90% rise time is about 2.5 ms the increase is less than 10%/Therefore itseems electrical filtering te the et of IPSCs observed in our work ie-7 4e relatvely sma the- mg乙 glycinergic synapses made on bullfrog BCs, just as suggested for glycine receptors in the neuron of the inferior colliculus( Backus, Frech and Karushaar Zoology 2000). This assumptiom is strongly supported by a recent paper/(Frech Perez-Lebn, Wassle and Backus, J Neurophysiol 2001) performing on mouse nacrine cells They-found that,-op compact amacrine cells. the 10-90%6 iset time of glycinergic IPSCs is about 2.4 ms, while that of GABAergic IPSCs is much less(see pages 1632, 1637-1639 of the paper). They thought that the slow rise time is not due to process fitering, but may be due to synaptic properties, such as slow association rate constants of glycine receptors, slower transmitter exchange rate, which could be due to different morphological properties, different glial ensheathment, or different transmitter reuptake rate and sparse clustering of glycine receptors resulting in a more desynchronized activation of the glycine receptors. Additionally, Maple and Wu(J Physiol 1998) C L %amm上(r