A source Sound can be shared if its sample stream can be routed to two different processing modules. Two Sounds that start at different times are no longer generating the same sample stream, though, especially if you have EventExpressions in the parameter fields.
If your two Sounds are truly identical (including fixed, identical parameter values), then you could "cache" that Sound on the first playing and thereafter use the cached version. Or you could capture the sample stream in RAM with a MemoryWriter and play it back from RAM at later times. Or you could route the Sound into a DelayWithFeedback whose delay time is set to the duration between the desired playback times. If the parameters are fixed, you could also create a recording of that Sound and play that recording back at different times in your Timeline.