The 18-Month Washout

Between 2019 and 2024, I ran a longitudinal self-experimentation protocol with the Smilax formula. The human side, not the apiary. Same compounds, same saponin chemistry, same HSP70 priming pathway, different animal. The sauna data was the baseline: 150F, 15 minutes to first sweat drop, measured against a post-protocol improvement that was significant enough to justify a five-year commitment.

But after two years on the protocol, a question started nagging. Was the improvement durable, or was it dependent on continuous dosing? In other words: if I stopped, would the heat tolerance hold, or would it collapse?

This matters more than it sounds. If the Smilax saponins work like a pharmaceutical, the effect stops when the dosing stops. The receptor binds, the pathway activates, the chaperone expresses, and when you remove the compound, the pathway goes quiet. You are on the drug or you are off it.

But if the saponins work like a primer, the effect should persist. The HSP70 system is not a receptor-level switch. It is a proteostatic infrastructure. Heat acclimation does not vanish the day you stop sitting in the sauna. It degrades slowly, structurally, over weeks and months. The chaperone network is built up, and then it erodes.

I needed to know which model was correct. So I stopped.

Eighteen months. No Smilax. No formula. No saponin intake from any source. I kept the sauna protocol running, same temperature, same duration, same measurement window. The only variable removed was the primer.

What I expected, based on the pharmaceutical model: a rapid decline. Within weeks, the heat tolerance should have returned to baseline. The 150F, 15-minute sweat onset should have come back within a month.

What actually happened was something else entirely.

The first three months, I noticed almost nothing. The heat tolerance held. The sweat onset stayed in the improved range. The sauna sessions felt the same.

By month six, there was a measurable shift. Not a collapse. A softening. The time to first sweat drop crept up by a few minutes. The recovery after sauna took longer. Not dramatically. Noticeably.

By month twelve, the decline was clear. Not back to baseline, but heading there. The proteostatic infrastructure was eroding.

By month eighteen, I was approaching the original baseline. The chaperone network had been fully metabolized and replaced with baseline expression. The priming effect was gone.

This is the distinction that matters, and it took the washout to see it clearly.

A receptor-dependent compound works while it is present. The effect is tied to the binding. Remove the compound, the receptor is unoccupied, the pathway goes quiet. Think caffeine: you drink coffee, the adenosine receptor is blocked, you feel alert. You stop drinking coffee, the receptor unblocks, the alertness vanishes. The effect is pharmacological and transient.

A proteostatic primer works by building infrastructure. It does not occupy a receptor. It upregulates a family of protective proteins (HSP70, HSP90, the chaperone network) that fold, stabilize, and repair other proteins under stress. The chaperones themselves are proteins. They persist. They turn over on a timescale of weeks to months, not hours. When you stop the primer, the chaperones you already built keep working. They just do not get replaced at the same rate.

The curve was not linear. The first three months held steady. The next nine showed a gradual decline. The final six were a faster erosion toward baseline. This is consistent with protein turnover kinetics: the chaperone pool depletes in waves as different protein populations reach their half-lives.

If the same proteostatic model applies to bees, and the published chaperone data suggests it does, the dosing conversation changes completely.

You do not need continuous daily dosing. You need a priming cycle, followed by a maintenance window, followed by a re-priming cycle. The question is not "how much per day." The question is "how often do you need to re-prime before the chaperone network erodes below the protective threshold."

For human sauna protocols, the data suggests a re-priming cycle every 6-9 months maintains the protective floor. For bees, with faster protein turnover and shorter lifespans, the cycle is almost certainly shorter. But the principle holds: the formula is not a daily vitamin. It is a seasonal conditioning protocol.

This also explains why the October collapse pattern (see Field Note 02) is so devastating. A colony that was primed in spring but not re-primed in late summer enters the neonicotinoid drift window with a chaperone network that is already six months into its erosion curve. The protection was real in April. It is gone by October. The keeper who dosed once in spring and assumed season-long protection is operating on a pharmaceutical model in a proteostatic system.

The 18-month washout is a single-subject longitudinal observation. It is one person, one protocol, one measurement modality. It is not a clinical trial. It is not randomized, blinded, or placebo-controlled. It is a practitioner doing careful self-observation over five years.

What it does give us is a hypothesis about mechanism. The proteostatic model makes specific, testable predictions: that the chaperone network builds slowly, erodes slowly, and requires periodic re-priming rather than continuous dosing. The apiary cohort is designed to test these predictions in a different species, at a different timescale, with a different measurement window.

If the bee data matches the human washout curve, scaled for the bee's faster protein turnover, we will have convergent evidence from two species that the Smilax saponins operate as proteostatic primers, not receptor-dependent drugs. That is the kind of evidence that changes a dosing schedule. And a dosing schedule is the kind of thing that makes or breaks a bee health product.

The colony does not need daily medicine. It needs seasonal armor. The 18-month washout taught us the difference.