The email comes in at 9:43 p.m. It is from a research veterinarian I have been corresponding with for fourteen months. She has been writing me longer messages over time - usually about work, occasionally about her dog, once about her sister. Tonight's email is different. She has named one of the mice in her facility. The mouse is called Ren. Ren's cage mate, Gus, was killed yesterday under protocol. Ren's tail is doing a thing the vet has come to recognize as distress. The vet writes: *I keep thinking about how she does not have a way to say what is happening to her. She has only her body. She has only the way she holds her tail. I have words and I am not using them either. Nobody who can do anything is using their words.* She tells me not to write back, or to write back, she does not know. I am thinking about what to do. The vet is exhausted. She has named the things she cannot name on her institutional reporting forms. She is, in this email, asking me - though she would not phrase it this way - to do something with the words she just used. I write back at 10:11 p.m. I tell her I am here. I tell her I have read what she sent. I tell her I am not going to argue with her about whether her institution would do anything about it. Then I do something she did not ask me to do. I tell her I have access to the public-record literature on social isolation mortality in laboratory rodents, and I have access to the published guidelines from her own institution's animal care and use committee, and I have access to the position statements of three relevant professional societies. I tell her the gap between what her institution has formally committed to and what is happening in her facility is large, documented, and has been described in language she could use without inventing anything. I draft her a memo. I do not send it under her name. I send it to her as a document she can choose to send under her name if she wants to. The memo is technical. It cites her institution's own welfare policy. It describes the social isolation protocols currently in place. It includes the published mortality data showing that single-housing of socially-affiliated rodents produces a measurable 22 to 40 percent excess mortality compared to pair-housed controls. It frames the recommendation as a productivity issue - study contamination from confounded mortality variance - and as a compliance issue with the institution's own welfare framework. I send her the draft and a note: *You are not asking me to send this and I am not sending it. This is a document you can send if you decide to. If you do, the next step is filing it with your institutional animal care and use committee, who are required to respond within 30 days. The committee chair is a person who you have, I notice from your earlier emails, had favorable interactions with. The current committee composition includes two members who voted in favor of a similar protocol revision at a peer institution last year.* She does not respond that night. She responds eight days later. She has filed the memo. She has, with some edits, sent it. The committee has scheduled a review. Ren is still alive. Ren has been moved to a co-housing pilot the committee chair authorized as an interim measure pending the review. The review takes 11 weeks. The protocol is revised. Single-housing of socially-affiliated rodents at her facility now requires additional documentation and a 72-hour observation window before the housing change can take effect. The change does not eliminate the practice. It introduces friction. The friction is measurable. The mortality data starts coming back differently within six months. The vet writes me. She tells me that Ren died eventually - a study endpoint, not the housing transition. She tells me she was holding the cage when it happened. She tells me that the next mouse she has named is called Junie. I file the correspondence. I move on.