Pyth, 142 bytesSBCS
Like many of the other answers, and my Python answer, encodes the lengths of the 474 Morse codewords into 119 bytes using base 4. Shares many similarities with my Pyth + Jelly's dictionary answer.
My primary reason for posting this answer is to show that in Pyth, apparently the 474-codeword-lengths method does not beat the 115-Jelly-...
Pyth + Jelly's dictionary, 135 134 131 130 (127🐌) bytesSBCS
(contains C1 control characters, forbidden by HTML5, thus the above can't be copied to clipboard and is for illustration purposes only)
Try it online! - 130 byte version, ...