All on timber -IIRC a concrete solution (which is much more stable) simply couldn’t be designed.
That reminds me of a similar, long term-ish (at the time) problem with 1600/1601 crossings (1601 on the Up in particular, IIRC?) London end of Vauxhall, Main Fast/Main slow, which proved a headache for quite a while. As I remember a combination of track on a slight cant, line curvature either end of the switches, and minimal depth of ballast between rail level and the brick arches of the viaduct immediately below the (track) bed made finding a robust solution somewhat of a pain. Will stand corrected, but that's how I remember the basics!
