I need to procedure a assay with KIT Elisa to assess levels of testosterone and oxytocin in the serum of mice, but I didn't found reference to sample dilution.
It depends of your kit, more specifically of the antibody and the standards used. I usually do a trial with various serial dilutions (for example: 1:25, 1:50, 1:100, 1:200) to find out which dilutions falls in the middle of the standard curve
First thing to know if possible is the basal level of these hormones in mice serum, then you should check if this level is compatible with the sensitivity/limit of detection of the ELISA kit you want to use or choose one in accordance with the level of hormones to assay. And then dry several serial dilutions as suggested by Jean-Loup Rault.
Besides, you should be very careful about the sampling for oxytocin measurement since it is usually recommended to draw blood on specific tubes containing an anti-proteases such as aprotinin and to pre-chill those tubes because of the delicateness feature of this nonapeptide and the risk of degradation.
Also, for this kind of biological matrix (serum or plasma), it is strongly recommended to perform an extraction step (on C18 columns in general) before running your samples in the ELISA. Indded, a lot of interferences have been shown in ELISA when samples are not extracted before, leading to some discrepancies and bias in the assay.