Counting a Population That Links: Incubation and Spread of the Morphogenic Disease Among the Founders, Reconstructed from Dominion and Federation Records to the Treaty of Bajor
Abstract
The morphogenic disease reached the Federation's notice in 2375 through one patient on one station, and ended with a cure that reached the Founders at the close of the Dominion War and was carried into the Great Link after it. Almost nothing about its course among the Founders was ever measured by anyone outside the Dominion. Writing in 2376, after the Treaty of Bajor, we reconstruct what can be said about its incubation and spread from 23 Dominion medical documents captured at five installations, together with a single Starfleet Medical patient file. We count a "case" as a Founder recorded as showing flaking or loss of form (19 in all) and a "contact" as a dated linking with an affected Founder (11 cases have both dates). The interval from the latest recorded contact to first symptoms had a median of 6 months (range 3 to 11), and leaving out any one installation moved the median only between 6 and 7 months. Recorded cases rose by about 1.5 times per quarter from the second quarter of 2374 to the second quarter of 2375, a doubling time near five months, with an interval from under three months to over two years. We name one alternative, dose accumulation across repeated linkings, and give our reasons for preferring the simpler estimate. The estimates are ranges, because the data were held by one belligerent and the pathogen was made by another party.
1. Introduction
The outbreak registry of the Federation Interspecies Medical Exchange was opened on Denobula in 2312, as a place to file diseases that crossed from one species to another (and a few that crossed back), and its rules of entry have always assumed that a case is a body. The assumption has an old source. Epidemiology took its working form in the investigation of Earth's great epidemics, where an unseen agent was inferred from the pattern of the sick, and that inference works only if a population is a set of separate persons, each with a date of exposure, a date of onset and a short list of contacts.
Since then the method has been stretched a long way. Denobulan, Vulcan and Andorian outbreaks have all been fitted into it (with some effort), and the registry has handled a few awkward cases of its own (Ferrante-Obi, 2372), but a population of Founders does not fit it at all.
Where the Founders gather in the Great Link there is no clean line between one patient and many, and a linking is at once a contact, an exposure and a kind of merger. When the morphogenic disease was first recognised, no one had yet asked how the method might serve a population that "merges". The Dominion recognised the disease first (through Vorta physicians). The Federation recognised it in 2375, through one patient living on one station, whose illness was traced in later accounts to an exposure during a visit to Starfleet Medical in 2372 and to a body that the Federation does not officially acknowledge (Vessenden, 2375). The cure, recovered from that body's own agent and tested in haste, reached the Founder who led the Dominion's forces in the Alpha Quadrant in the last hours of the war, and was carried into the Great Link only after the Treaty of Bajor was signed in 2375 (Orsolya, 2375; Federation Council Legal Research Office, 2376).
We write in 2376, after the Treaty, and every source we use predates it.
Our aim is modest: to say how long the disease likely took to show itself after a linking, how fast it spread through the recorded Founders, and how little the record allows us to say about either (less than we would like).
2. Sources and Method
Three bodies of record bear on the question. The largest is a set of 23 Dominion medical documents, mostly Vorta assessments of the condition of named or designated Founders, recovered from five installations and catalogued by Starfleet Intelligence (Starfleet Intelligence, 2376). Most assessments concern Founders in the Great Link itself, and reached the Alpha Quadrant installations as copies forwarded by Dominion couriers. The documents do not say by what route after 2374, when the Alpha Quadrant forces were cut off, and transit time is one more source of error in the dates. Who wrote the assessments, and for whom, has been described from the same captures (Sarrafian, 2376). The second body of record is the Starfleet Medical file on the one Federation patient (Vessenden, 2375), supported by working notes on the cure (Orsolya, 2375). The third is older and was written without the disease in view: a study of the cellular organisation of the changeling form, from which we take our baseline for what "flaking" departs from (Bellweather, 2374).
We counted "cases" as Founders recorded in a document as showing flaking of the outer form or loss of form, and "contacts" as any dated linking with a Founder already recorded as affected. Neither definition is precise (we concede as much), but the documents allow no better. Nineteen Founders meet the first definition. Eleven of them also have a dated contact and a dated first symptom, with dates good to about a month (and a month is generous for some documents) once the Dominion's calendar is converted, a conversion that Starfleet Intelligence puts at an accuracy of two weeks or so. We follow the practice of the literature on colonial organisms in treating the linking itself, not the individual, as the event to be counted (Teveral, 2373).
For incubation we report the median, the quartiles and the range of the interval from latest recorded contact to first symptoms, with an interval for the median built from order statistics (since eleven values do not support a normal approximation), and we recompute the median leaving out each installation in turn. For spread we count first records of symptoms by calendar quarter and fit a log-linear Poisson trend, reporting a Wald interval for its slope that is only approximate at these counts. We attach no p-values, since eleven intervals and five counts are too few to carry a test.
3. Incubation After a Linking
For the 11 Founders with both dates, the interval from the latest recorded contact to first symptoms had a median of 6 months, with quartiles at 5 and 7.5 and a range of 3 to 11 (Table 1). An interval for the median built from order statistics runs from 5 to 8 months and has a coverage of about 93%. We took this as our estimate of incubation, and we would call 4 to 9 months a fair statement of what a single Founder might expect.
We were surprised to find how well the crude definitions held up under a check that could have failed. The 11 intervals come from the five installations in groups of 3, 3, 2, 2 and 1, and leaving out each installation in turn leaves a median between 6 and 7 months. Single-installation medians run from 5.5 to 7. The Vorta who wrote the documents shared no form and no reason to coordinate.
The plainest alternative to our estimate is that the clock starts earlier. Founders in the same "circle" link repeatedly, and if each linking adds to a "dose", the interval would run from the earliest recorded contact, giving a median of 7 months and a range of 5 to 14. We think the latest-contact estimate more likely, for a reason the sample can show. Four Founders have a single recorded contact, so the two definitions coincide for them, and their median is 6.5 months. The other seven, with several contacts, have a median of 6 months from the latest contact but 10 months from the earliest. If later linkings added nothing, the earliest-contact figure would stay close to the single-contact group, and it does not. The latest-contact figure does. Two things favour the shorter figure regardless of biology: a latest contact is by definition the nearest to onset, and a record ending in June 2375, with cases still rising, cannot contain the longest intervals from recent contacts. Both push the estimate down, so 6 months is a lower middle and not a centre.
An earlier Starfleet Intelligence summary put the period at nine to twelve months (Starfleet Intelligence, 2375). Its authors were evidently unaware of the circles' separate schedules, and took the gap between the first and last cases in a circle as the incubation period, so that spread and incubation were folded into one number. Nobody could have separated the two then, since the contact dates that do so were not yet in Federation hands.
4. Spread Through the Recorded Founders
The 19 recorded cases have first symptoms dated to the five quarters from April 2374 to June 2375 (Table 2). No case in the captured set is dated earlier, and none is dated later, since the latest document in the set is from June 2375. The counts rise from 1 in the second quarter of 2374 to 7 in the second quarter of 2375.
A log-linear fit gives a slope of 0.42 per quarter, with a 95% interval of 0.07 to 0.78. That is a rise of about 1.5 times per quarter (interval 1.07 to 2.18), and a doubling time of 4.9 months. The doubling time runs from 2.7 months to about 30, because five small counts cannot fix a slope tightly (and we do not pretend otherwise). If we use the median incubation of 6 months as a rough "generation time", the recorded cases multiplied about 2.3 times per incubation period, with an interval from 1.1 to 4.8. An incubation anywhere from 4 to 9 months widens that to roughly 1.1 to 10.
Two cautions apply.
The installations that yielded documents were taken at different points in the war, and each holds only what had reached it before it fell, so the rise in recorded cases may partly be a rise in recorded documents (and the counts alone cannot tell the two apart). The documents themselves hold steady at four or five per quarter, and cases per document run 0.25, 0.60, 0.40, 1.20 and 1.75, rising from 0.44 in the first two quarters (4 cases in 9 documents) to 1.44 in the last two (13 in 9). That is why we think the growth is likely real nonetheless, though the lower end of the interval, close to no growth, cannot be ruled out.
Another concerns the unit. A case count says how many Founders were recorded and nothing about the share of the Link affected, because the Link's population is unknown to us.
5. Seeding and the Missing Beginning
If the incubation estimate is applied backwards to the earliest case, it dates the infection of that Founder. The first recorded symptoms fall between April and June 2374, so an incubation of 4 to 9 months puts the infection between July 2373 and February 2374. Because the earliest recorded case need not be the earliest case, this is a latest date for the seeding of the Link, and the Federation record (below) puts the seeding earlier still.
That record bears on the window directly. The exposure on Earth is placed in 2372, and the patient's illness first appears in the station's infirmary record in 2375, an interval of about three years. We do not pool this interval with the 11 above, since the exposure was not a linking and its dose, strain and route are unrecorded. The patient's linking with the Great Link is itself on the Federation record: it took place on the Founders' homeworld in late 2372, within months of the exposure on Earth, and the patient then spent some months held in a solid form (Vessenden, 2375). If that linking seeded the Link, the seeding came more than a year before the latest date our incubation estimate allows, and the Dominion record misses the first generations of cases.
The window has one further implication. If the Link was seeded in late 2372 and the first recorded symptoms fall in the second quarter of 2374, the disease ran for something like a year and a half with no case that the captured documents record (the Vorta may not have seen it, or not written it down, or their documents were never captured). The growth we measure began well after the outbreak did, and the early counts of 1, 3 and 2 carry the least weight.
6. Discussion
Our estimate is that the disease likely took about six months to show itself in a Founder after the latest linking with an affected one, with 4 to 9 months a fair range, and that recorded cases rose by around half again each quarter, a doubling time near five months. We are fairly sure of the first claim, though its figure is a lower middle, and less sure of the second, which could change with a single further document.
That first claim rests on the 11 intervals and on one comparison. The alternative, measuring from the earliest contact, moves the figure to about seven months and the range to 5 to 14. We prefer the shorter estimate because the Founders with one recorded contact, for whom nothing is in dispute, agree with it. A dose effect that the sample is too small to show cannot be excluded, and the longer figure may be the better guide for a Founder long linked with an affected circle.
Of the two, the second claim is weaker. The interval on the rate is wide, and the unit is a recorded Founder, not a share of the Link. What the counts probably support is that the disease was increasing in the Dominion's own records through the spring of 2375, at a pace implying anything from a small rise to several-fold growth per generation of cases. That fits the haste with which the cure was used in the final months of the war. It records what the Dominion's physicians wrote down, which is not a measurement of the Link.
Part of the point is method. A population that links offers no clean individuals to count (that was the difficulty from the start) but does offer dated linkings, and these gave a median that does not depend on its longest intervals, which is all a median of eleven can promise.
7. Limitations of the Record
Thin is the best word for the record, and it is thin for reasons that cannot be mended. All the data were held by one belligerent. The Founders distrust solids and let the Vorta examine them only on their own terms, so the 23 documents describe a population observed rarely and reluctantly, by physicians with every reason to report favourably. The documents survive because installations were overrun, so they over-represent the units that lost, and the route by which assessments of the Link reached those installations is unknown to us.
A third party made the pathogen (its makers have not published, and we did not expect them to), and the Federation's side of the record begins only after the disease was known. Starfleet Medical holds no document on the engineering of the agent, its strain or its intended "dose" (if it had one), and its file for the one patient says little about the illness before 2375 (Vessenden, 2375). We cannot say how the pathogen was meant to behave, so we cannot say how far the observed course departs from the intended one.
Small samples carry their own limits. Eleven intervals support medians but no covariates, and the 8 cases without contact dates may differ from the 11 that have them. The case definition is the Vorta's. A Founder who held a solid form while carrying the disease would not appear, nor one who died before an assessment was written. Because the record ends in June 2375 while cases were rising, the longest intervals are the ones most likely to be missing, and the incubation estimate is biased short for that reason. We do not know the size of the Link, so no rate can be given. The quarterly series has five points and ends before the cure.
Dating relies on the Dominion's calendar, converted to within about two weeks, and on our decision to count the linking as the contact. Every estimate here is a range, and the point values are only its middle.
This paper extends the registry's work on populations that do not count as individuals (Ferrante-Obi, 2372; Teveral, 2373). Readers who want the documents should start from the captured-materials catalogue (Starfleet Intelligence, 2376), whose notes on the conversion of dates are the best guide to the errors we have only summarised here. On the general problem of counting organisms without clear boundaries, Teveral (2373) is the best entry point, and the registry's own case files on Denobula hold the rest.
References
- Starfleet Intelligence (2376). Vorta medical assessments of the condition of the Founders, recovered from Dominion installations in 2375, translated with catalogue. Starfleet Intelligence Captured Materials Series, CMS-76-9, items 1–23.
- Vessenden, A. (2375). Case file on a morphogenic illness in a changeling patient, compiled from the infirmary record of Deep Space 9. Starfleet Medical Records Archive, SM-P-75-14.
- Orsolya, K. (2375). Working notes on the recovery and testing of a morphogenic cure. Starfleet Medical Records Archive, SM-X-75-08.
- Bellweather, N. (2374). Cellular organisation of the changeling form, observations from one long-term patient. Federation Journal of Xenomedicine, 62(2), 140–167.
- Teveral, M. (2373). Counting cases in colonial and syncytial organisms. Comparative Xenobiology Review, 14(1), 30–52.
- Ferrante-Obi, L. (2372). Outbreaks that do not fit the individual-case model, a review of registry experience. Starfleet Medical Journal, 95(3), 120–141.
- Federation Council Legal Research Office (2376). The Treaty of Bajor, text, annexes and signing record. Federation Council Legal Research Office Reports, FCLRO-76-04.
- Sarrafian, D. (2376). The Dominion medical establishment as seen from captured installations. Journal of Dominion War Studies, 1(2), 56–84.
- Starfleet Intelligence (2375). Intelligence summary on the health of the Founders and the Dominion response. Starfleet Tactical Analysis Reports, TAR-75-31.
Open in Uncited Press →