I think you're confused about how the capacitors are working here. First, C1 gains charge Q on its plate. Next, C1, still with charge Q on its plate, is hooked up to C2, a capacitor with zero charge. You need to realize that no matter the distribution of charge on the two capacitors' plates, the sum cannot exceed or undershoot how much charge was initially put on C1. Otherwise, you have created or destroyed charge.
Since the capacitors are in parallel, you know they share the same voltage, so you can calculate how much charge is on C1 after the connection by using 35V and prior to the connection by using 50V. Using total charge in the system (initial charge on C1) and charge on C1 after the connection, you can subtract the two to find how much charge must be on C2. Finally, using the charge-voltage relationship of a capacitor (q = cv), you can calculate the capacitance of C2.
Using -15V seems really off. The delta V is usually written V and stands for voltage across the capacitor. It's not "change in voltage" as in "I went from 50 to 35 volts. Therefore, -15"