A little more information to add to what r norman posted:
As explained, for a response that can be fitted with the sum of two
f(t) = a exp(-t/t1) + b exp(-t/t2)
where t1 and t2 are the time constants and a and b are the initial
amplitudes of each component of the response at time t=0.
In some instances, rather than just quote both of the time constants
and their respective initial amplitudes, a single 'weighed' decay
constant is reported. Some times this is because a comparison is
being made between different synapses that may or may not have single
To calculate the 'weighted' mean time constant, the individual
initial amplitudes for each time constant are expressed as a fraction
of the total initial amplitude. Lets call these fractions A and B,
such that A = a / (a+b) and B = b / (a+b)
Lets then call the 'weighted' mean decay constant T, and this is given by:
T = A * t1 + B * t2
So you see, you get a indication of the exponential decay by
'weighting' the individual time constants according to the initial
amplitude, to get a mean time constant. This is reasonable, given
that the initial amplitude and the time constant of each component
will determine the proportional contribution to the exponential decay
that is being described.
I hope this helps.