The Unsolved Mystery of How Viruses Spread – and Why Germ Theory Isn’t the Whole Answer

For over a century scientists have struggled to prove viral contagion. There are gaping holes in the current model of viral transmission but these holes are not a reason to throw out the concept of a virus. We need a new model that fits the evidence. Here I set out the evidence for and against the current model and suggest a way to reconcile the apparent contradictions. 

Evidence for viral contagion 

1. Viral genetic material turns up in clusters of sick people. The sequences match. They change over time with new mutations in consistent ways without reverting. Even though testing is not perfect, people with positive tests are far more likely to be sick than not.
2. Viruses have been well described. Even if isolation methods are not flawless, electron microscopy and crystallography have shown fine-grained details including the shape of structures like the surface of the spike protein.
3. At high doses, transmission works. Human challenge trials demonstrate that viral exposure can cause illness when the dose is high enough.
4. Genetic tracing of viruses during outbreaks shows that distinct lineages spread from person to person in predictable clusters, confirming person-to-person transmission.

This evidence explains the mechanism of viral infection. But it does not explain the timing of the waves of infection that are characteristic of many viruses like influenza and COVID-19.

Where the traditional model fails

1. Normal-dose challenge trials often fail. The evidence here is strong: under experimental conditions, exposure frequently does not result in illness. A recent study confirmed this again.
2. Hospital-acquired infections peak at the same time as cases in the community. If spread were primarily driven by close contact, we would expect a lag, as community infections peak then admissions then within hospital infections. But the expected lag does not occur. In fact, hospital-acquired infections peak before the admissions to the hospital.  
3. Waves occur with seasonal regularity. Epidemic peaks in the UK often occur with peak deaths at predictable times of year before falling away for a time:
   • Every early January
   • Often early April
   • Sometimes in July
   • Every late October

In the USA, the southern states saw more of a summer wave during Covid and wastewater analysis for viral load shows a repeating pattern that continues to this day. The January 2022 Omicron wave (which occurred after a huge Christmas vaccine booster campaign) was exceptionally sharp in its rise and fall and followed by a deficit. However, the timing of peaks and troughs is predictable, with a low every spring and a peak every summer and winter.

This is not explained by school terms or behavioural cycles alone. Lockdowns were a convincing (albeit unethical and very harmful) experiment that proved these things do not interrupt transmission. The steepness of the slopes and the timing of the peaks remained as would be predicted in the absence of lockdowns. The vaccinations did seem to impact the steepness of the January 2022 waves and the baseline levels of virus between waves from 2022, by increasing both.

The illusion of viral timing

It is often claimed that respiratory virus waves are driven by viral evolution or viral interference. In this view, waves happen when new variants arise that can evade existing immunity, which increases transmissibility and allows the virus to reinfect previously exposed populations. This idea has been central to the interpretation of influenza seasonality and SARS‑CoV‑2 waves, where ‘variants of concern’ with altered transmissibility or immune escape properties have dominated successive waves of infection worldwide.

If that were true, the pattern of waves should look a certain way:

• The spacing between waves would vary depending on when mutations appear or viruses outcompete each other.
• Some waves would be tightly packed, others widely spaced.
• There would be frequent overlaps of waves when mutations happened in quick succession
• Faster-mutating viruses would cause more frequent waves.

This is not what happens.

The reality is much more regular and much harder to explain by mutation or interference alone.

1. The timing is too consistent
In the UK and many other temperate countries, waves have arrived roughly every 13 weeks, across multiple years and virus types. This rhythm has held steady through variant shifts, travel restrictions and mass behavioural changes. Random events like mutation and competition do not produce this kind of precision.

2. Submariners are not protected from infections

According to the conventional virus-centric model, a respiratory virus wave ends when population-level immunity is achieved and subsequent waves occur only with new mutations. The rate at which mutations can occur in a closed environment will be a fraction of what can occur globally.

From this logic, we can infer what should happen in a sealed environment like a submarine.

1. The crew is a closed, healthy population.
2. Any circulating respiratory viruses will cause infections at the start of deployment. 
3. Once initial transmission occurs, everyone who is susceptible gets infected.
4. The virus runs its course.
5. The crew should then be universally immune to the strains that are aboard.
6. No further illness should occur unless a new virus or variant is introduced.

But this is not what happens. The rate of acute respiratory infections was identical in submariners and those on the surface. In real-world submarine missions, respiratory infections remain common across the full duration of patrols. One study measured medical consultations from a centralised onboard medical system covering 240 submarine patrols lasting at least 10 days between 1997 and 2000, encompassing 13,441 male crew members (1,389 officers and 11,952 enlisted personnel) and over two million person-days at sea. Unlike for other illnesses, crew and officers and young and old had similar rates of respiratory illness. While they remarked on lower overall illness after 60 days at sea they did not claim there was any difference in the distribution of illness in this period. Unfortunately, they did not set out to measure this so we have no exact data, only the implication.

Despite minimal viral input and little opportunity for mutation or new exposure, new cases continue to arise. This directly undermines the idea that viral mutation is the rate-limiting step in transmission. New viruses and viral evolution are not driving these infections but changing host susceptibility may be.

Submariners are exposed to high levels of sewage derived bacteria which are maintained at constant levels in the air for weeks on end. These bacteria are detectable in the nasopharynx and in other circumstances are pathogenic and yet the rates of illness, both respiratory and gastrointestinal, remain as expected.

3. Faster mutation does not speed things up

In the 24 months before December 2021, pre-Omicron SARS-CoV-2 accumulated around 20 mutations per year. From January 2022 to mid-2025, Omicron accumulated approximately 25 per year. Despite this 25% increase in mutation speed, the waves kept arriving on the same seasonal schedule. 

4. Influenza mutates faster still and behaves the same

Influenza A mutates more rapidly than SARS-CoV-2, yet it too peaks once per season at a maximum. If mutation controlled wave timing, influenza should be constantly surging. But it too peaks at the same times of year as the slow to mutate SARS-CoV-2.

The trajectory of influenza and RSV is thought to look like this:

However, this prevalence estimate is based on testing of people who are sick enough to seek medical care and does not reflect the whole picture.

Conventional surveillance testing is focused on the sick and suggests that influenza only has a significant presence in autumn and winter. However, when testing is expanded to less sick people as happened in 2009 there is clear evidence of a peak in influenza each season.

Similarly, WHO’s global RSV surveillance strategy explicitly dropped the fever requirement for testing, recognising that around half of RSV cases, especially among children and the elderly, presented without fever and were systematically excluded from testing. While formal results are pending, the change reflects a recognition that traditional surveillance has masked the true scope and timing of RSV circulation.

5. Hospitals are full of virus aerosols year-round but infections still peak and fall

Air samples from hospitals consistently show SARS-CoV-2 present, even between waves. If exposure alone drove infection, we would expect ongoing hospital-acquired spread. But in-hospital infections rise and fall in synchrony with the community, including during troughs. The virus is still present in hospital air but the susceptibility is absent.

6. Spread skips regions

Each variant of SARS-CoV-2 spread country to country following the same seasonal susceptibility patterns as influenza. Large regions were skipped entirely in early waves e.g. Eastern Europe in spring 2020 and UK, Portugal, Ireland (i.e., the west of Western Europe) in spring 2021. If attack alone caused spread, this would not happen. But if susceptibility rises regionally, the virus already present in the air becomes dominant when the window opens. This would also explain why variants ‘spread globally’ at the same rate each time. Virus was also present in regions with no susceptibility and never took off, for example, all of South East Asia was not susceptible to Covid without any exceptions until all of South East Asia became susceptible without exceptions in early 2022.

7. Infections are determined by what is in the air at the time of susceptibility including coinfections

Often more than one respiratory virus is detected in an infection. So the idea that one virus has successfully outcompeted all others across a population does not hold up – if they cannot outcompete in an individual how could they in a population? When susceptibility rises, people inhale whatever is already in the air. That may be one virus or several. The virus that dominates is simply the one most widespread at that time and place.

We have seen this play out in ‘W-shaped‘ epidemic curves, where one virus briefly overtakes another mid-wave. Often these are sequential waves, but occasionally a Covid surge dips as influenza rises, or vice versa, then rebounds within one season. This shows that competition between viruses is real. But the overall wave, the rise, peak, and fall, stays on schedule. That means the wave is not being driven by the virus. It is being driven by susceptibility. The competition only shapes the subwaves of individual viruses inside it.

Why susceptibility fits better

If we assume viruses are circulating most of the time, but that illness only occurs when susceptibility rises in a portion of the population, then the pattern starts to make sense. The virus that dominates a wave is simply the one most prevalent in the air when that window opens. Regional synchrony, global emergence patterns and stable spacing between waves all follow logically from this.

Susceptibility explains:

• Why waves are regular
• Why whole regions can be skipped and hit later
• Why hospital infections synchronise with the community
• Why faster mutation does not change the tempo

Mutation and interference determine which virus dominates a wave but not when the wave arrives. Whatever is controlling when people get sick, it is not the virus.

The susceptibility model

The model that makes sense of these observations is one where infectious agents are necessary but not sufficient. The timing of illness must be due to something else.

There must be a third factor that:

1. Peaks once each season (always in autumn and winter and not always in spring and summer)
2. Affects only a fraction of the population each season
3. Can spare certain regions entirely in any given wave
4. Is capable of synchronising illness peaks across hospitals and communities

We do not know what this factor is but the data suggest it exists.

The evidence excludes certain possibilities:

• Human behavioural changes like school terms made no difference to peaks in lockdown.
• Temperature and humidity are important for viral stability, but do not have four seasonal peaks.
• Pollen, spores or volatile organic compounds do vary over the year, but not with the precision or global consistency needed.
• Other environmental factors like electromagnetic or space weather effects have cyclical differences but these track annually, not quarterly.

Any of these may play a role, but none independently explain the pattern. What might?

The immune clock

Our immune systems change with the seasons. Gene expression studies show clear and consistent shifts in immune pathways across the year – with January and July as polar opposites and transitional patterns in April and October. These changes are not subtle – nearly a quarter of genes are affected. The pattern is the opposite in Europe to Australia. In some cases, the winter and summer immune profiles are as different as those seen in entirely distinct disease states. These shifts in immunity profiles are reflected by changes to both the biome and susceptibility to pathogens e.g. COPD patients also show changes in their respiratory microbiome across different months of the year and influenza and coronavirus waves surge at predictable times as the immune profile shifts.

Susceptibility and dose

The two forces of viral dose and host susceptibility interact. In controlled human challenge trials, where viral doses are carefully administered and often far exceed natural exposures, symptomatic infection is induced in a higher proportion than seen in a natural wave. This shows that at high doses, exposure can overwhelm even a robust immune state, but that is the exception, not the rule. In real-world settings, exposure is often low from a few aerosols in the air, not a bolus of virus. In these low-dose exposures, host susceptibility becomes the dominant factor. If the immune system is in a defensive, resilient state, exposure may pass unnoticed. If susceptibility is high, due to seasonal, environmental or physiological factors, even a small viral input may be enough to trigger illness. 

Whatever the factors are that lead to susceptibility they must overcome the mucus barrier of the respiratory tract which is normally impenetrable to viruses.

Is this hypothesis falsifiable?

There are three main explanations for the seasonal cycling seen in human immune gene expression:

1. The viral mutation model. Immune cycling is downstream of viral exposure. Seasonal waves of infection, with everyone exposed to airborne viruses, drive immune activation, while gene expression shifts reflect that exposure.
2. Innate biological rhythm. Each person’s immune system runs on a built‑in annual clock, independent of environment or exposure.
3. Environmental entrainment. Immune function responds gradually to sustained environmental inputs – atmospheric, electro-magnetic or otherwise – which vary by season and location.

These explanations make different predictions. In particular, they differ in how quickly immune changes should appear when the environment or viral exposure changes.

If the rhythm is innate then nothing should change it. If immune cycling were primarily virus‑driven, then removal from seasonal viral exposure should produce rapid and marked immune changes. Conversely, if immune cycling reflects slow environmental entrainment, changes should accumulate gradually, with delayed recovery on return.

The fact that people in Australia have the opposite cycle of immune gene expression indicates that the rhythm is not innate.

The NASA twins study

The NASA Twins study provides a unique natural experiment. Astronaut Scott Kelly spent 340 days aboard the International Space Station (March 2015-March 2016), while his identical twin Mark remained on Earth as a control, exposed to normal seasonal cues and circulating viruses.

Scott was removed from Earth’s atmosphere, magnetic environment and from seasonal respiratory virus waves.

• Immune gene expression changes did not change immediately after launch.
• Instead, the number of differentially expressed genes was modest in the first six months and nearly six‑fold higher in months six to 12.
• On return to Earth in early March 2016, over 90% of gene expression changes reversed only slowly, over several months.
• 811 genes, including immune and DNA repair genes, remained altered six months after return.

If immune cycling were primarily a response to viral exposure, the most dramatic changes should have occurred early in the mission or early on return, when seasonal viral input was abruptly removed or replaced. They did not. The immune system did not react sharply either on entering space or on returning. Instead, immune modulation accumulated gradually, suggesting a lagged adaptation to a sustained environmental change, not a rapid response to missing pathogens. These observations argue against both an innate immune rhythm and virus exposure as the primary driver. They support a third explanation: environmental entrainment. 

Further tests

This interpretation leads to specific, testable predictions:

• Astronauts or submariners returning to surface environments in different months would provide a natural test. If virus exposure alone were the key factor, immune markers should re-align with viral exposure with more rapid realignment during viral peaks. But if environmental entrainment is the dominant driver, then re-alignment would occur over weeks or months, regardless of viral exposure. Furthermore, testing for infections would demonstrate whether any infections were playing ‘catch up’ with viral waves that had passed in their absence or whether they were aligned with community infections.
• Antarctic researchers, overwintering in controlled environments with extreme photoperiod disruption, could help separate electro-magnetic exposures from microbial drivers.
• Passengers crossing hemispheres offer a natural test.
  • Viral mutation theory says viruses travel from Australia towards Europe. If virus exposure drives illness, those flying north to south (e.g. UK to Australia) should be at higher risk than those flying south to north (e.g. Australia to UK), who’ve already been exposed. But if susceptibility is driven by environmental entrainment, there should be no difference: both groups would be equally vulnerable when immune profiles are misaligned with local conditions. 
  • Furthermore, infection risk should peak for travellers in January and July, when immune gene expression between hemispheres is maximally misaligned, and be lowest in April and October, when profiles naturally align temporarily. If illness follows this pattern, then susceptibility, not the direction of viral flow, is what governs risk.

Possibilities going forward

A complex web of interactions between multiple environmental variables might possibly help explain the phenomenon. Together, they might create a susceptibility environment that emerges on a roughly 13-week cycle, even if no single input has that pattern.

However, it could well be something we simply have not measured e.g. a seasonal atmospheric phenomenon that is invisible to our current tools.

It has been 170 years since germ theory was accepted. Scientific ideas evolve and this one needs to. The evidence shows that viral attack alone does not explain all the observations and a more nuanced model is urgently needed. If we want to predict and prevent illness, we need to stop staring only at the virus and start listening to the human immune clock and the environment it inhabits.

Dr Clare Craig is a diagnostic pathologist and Co-Chair of the HART group. She is the author of Expired – Covid the untold story and Spiked: A shot in the dark.