I was calling writeMicroseconds() before and after just to cover all the bases to turn off the servo. Your previous comments only stated to include it prior to the attach(), but since that did not work, I included it after as well. I will remove it since it is not helping.
The values to stop the motors are the values that were determined while testing. Since they are hacked as I stated in my first post, I am pretty sure the resistors I used are not all identical and the variance is the reason for the different writeMicroseconds() values. These values do stop the motors and I have checked them over several minutes to verify.
The VarSpeedServo library came up during my research and I thought it was more appropriate than the standard Servo library for my continuous rotation servos. I will switch to the Servo library to see if it makes a difference. Thanks!!