You need to take into account the characteristic impedance of the relay itself, which in this case is 100 mOhm. 100 mOhm should not cause the behavior you are seeing.
The attenuation is more than likely due to attenuation due to your signal containing frequency components greater than the worst case channel bandwidth of the NI 2503, which is 35 MHz.
All signals, repetitive or not, can be broken down into fundamental frequency components per Fourier's Theorem. If your signal contains frequency components greater than 35 MHz, these components will be attenuated. Try inputting a sine wave to the relay/scope path and gradually increase the frequency. As it goes past 35 MHz, you should see it begin to attenuate. The rising edges of the spikes within your 16 pulses w
ill correspond to high frequency components which may be attenuated.