While its not surprising to see an immediate increase, it is strange that it would increase so sharply and then stop. As you say over the few months after the war began the casualties increased significantly so you would expect the rise here to increase constantly over those months to some degree. There was certainly space for it to do so as the percentage never really maxes out (I would say the max is below 100% given many family's will already have sons named after the father but still). There could have been some national push towards naming children for there fathers over one specific week but I think its more likely that there is some irregularity in how the data was collected which seems likely in war time.