Methodology: why dust, why qPCR, and how we read it

What this page is

ExplainmyERMI is an educational interpretation service for ERMI (Environmental Relative Moldiness Index) test results. The product is a 30+ page PDF report. It reads a laboratory-issued ERMI panel and gives a structured interpretation of the species pattern, in the context of the customer’s building and climate.

The report is meant as a starting point. It says what the species pattern is generally associated with, and it lists places to look and questions to raise with a qualified, independent professional. It is meant to help a customer prepare for an on-site inspection, not to replace one. It should not be used on its own to decide whether to remediate, to move, or to buy or sell a property, or to make any medical decision.

This page documents the choices behind that product, in eight sections:

1. Why we read settled dust, and why qPCR
2. How we read an ERMI result: the score, the Group, the Moisture Ladder, HERTSMI-2 and the building-context sections
3. A worked example from public data
4. How much a single result can move
5. The limitations
6. EPA’s position on ERMI
7. What the service is not
8. What you can check

It is written for readers who want the technical material that the homepage FAQ summarizes: clinicians, inspectors, industrial hygienists, regulators, and customers evaluating the service.

The figures on this page come from our own analysis of 1,790 dust samples from about 1,586 homes in the two HUD American Healthy Homes Surveys. EPA researchers led the mold analysis for those surveys. The two surveys are pooled without survey weights, so they make a large national sample, not a representative one.

1. Why dust-based testing

Dust is a time-integrated exposure proxy

Three processes shape the dust in a home. Deposition puts airborne particles onto surfaces. Resuspension lifts settled particles back into the air through foot traffic, vacuuming and HVAC airflow. Penetration brings outdoor particles in through the building envelope. Thatcher and Layton (1995), in a foundational paper on indoor particle dynamics, established that the dust composition observed in a home at any moment is the steady-state result of these three processes running continuously. Every footstep, every cooling cycle and every return-air pull puts some fraction of dust-bound particles back into the air people breathe.

This matters for two reasons:

• Dust composition is related to what has been in the air. What settles in floor dust is related to what has been in the air occupants breathe. The air carries it in small amounts, over the weeks or months the dust built up. The longer the dust has been collecting, the longer the period it represents. That period can include conditions that have since changed (Section 5).
• Air sampling captures one moment. A cassette or impactor running for a few minutes records the particles airborne during that window. Field studies have found that airborne mold counts in homes with reported mold problems are often no higher than in homes without (Strachan et al. 1990; Nevalainen et al. 1991). A short air sample can miss a problem that settled dust would register.

EPA researchers built ERMI on settled dust, not air. A settled-dust sample collects over months. An air sample is a spot reading, and it depends on what the air handler was doing in the hour before.

Fragments can outnumber spores

The most common objection to dust-based mold testing, and to qPCR-based testing in particular, concerns dead spores. Dust contains DNA from dead spores. The objection is that this DNA does not represent current exposure or ongoing biology in the home.

This objection assumes that intact, living spores are the only exposure that matters. The indoor-aerosol literature suggests otherwise.

Górny, Reponen, Willeke, Schmechel, Robine, Boissier and Grinshpun (2002) measured what contaminated surfaces release, in work led from the University of Cincinnati’s Center for Health-Related Aerosol Studies. They found that the number of fungal fragments released was consistently higher than the number of intact spores. That held under every condition tested, regardless of surface material, air velocity or vibration applied.

Fragments are particles smaller than a spore: pieces of hyphae and cell-wall debris, roughly 0.3 to 1.6 µm across. Intact spores are much larger, typically 4 to 5 µm for Stachybotrys chartarum and 2 to 3 µm for Aspergillus versicolor. Because fragments are smaller, they penetrate deeper into the respiratory tract.

Cho, Seo, Schmechel, Grinshpun and Reponen (2005) quantified fragment and spore release under controlled aerosolization. They reported:

• Stachybotrys chartarum released approximately 380 fragments per cubic centimeter, roughly 500 times as many fragments as intact spores. Stachybotrys spores are held in slimy heads and are intrinsically difficult to aerosolize. The fragments are not.
• Aspergillus versicolor released fragments and spores in comparable numbers.
• Respiratory deposition modeling using LUDEP 2.07 found that for Stachybotrys, fragments produce 230 to 250 times the respiratory deposition of spores. Fragments are smaller and more numerous, and the two effects compound.
• In the same model, the ratio of deposited fragments to deposited spores in the lower airways was four to five times higher for a 3-month-old infant than for an adult. This is a modelled ratio, not a measured health effect.

Brasel, Douglas, Wilson and Straus (2005) detected Stachybotrys macrocyclic trichothecene mycotoxins, by immunoassay, on airborne particles smaller than spores. This shows the toxins can travel on fragments. It does not by itself show how much anyone is exposed to, or what the health effect is.

Implications for the qPCR vs viability question

A spore does not need to be alive to carry DNA. qPCR therefore counts non-viable spores and DNA-bearing fragments of hyphae, which culture methods miss entirely. That matters because exposure does not depend on viability. Dead spores and fragments still carry allergens and, for some species, toxins.

qPCR does not count the smallest cell-wall debris directly. What it measures is how much of each species has been present, and those are the organisms that shed fragments. The fragment figures above are particle counts from chamber studies. An ERMI result is not a fragment count.

The “qPCR detects dead spore DNA” objection is therefore true, but less damaging than it sounds. For exposure, counting what culture cannot grow is a strength.

2. How we read an ERMI result

A laboratory ERMI report lists a concentration for each of 36 species and, usually, an ERMI score. The concentrations are in spore equivalents per milligram of dust. EPA’s papers call the same unit cell equivalents. The ExplainmyERMI report reads those 36 numbers in several separate ways. Each reading is described below: what it is, what it reports, and what it is generally associated with.

2.1 The ERMI score

The score is the sum of the log₁₀ concentrations of the 26 water-damage indicator species (the panel’s Group 1 list) minus the same sum for the 10 common-indoor species (its Group 2 list) (Vesper et al. 2007).

Two things follow from the subtraction. When both sets of species are elevated, the subtraction cancels them and the score can come out near zero. When the common-indoor species are nearly absent, there is little to subtract and the score can come out high, even if few indicator species are raised.

The scale also depends on how an undetected species is handled. The quartile boundaries usually quoted, about -4, 0 and 5, match the first survey’s data when an undetected species is entered as 0.5 spore equivalents per milligram, as that survey’s data file records it. Our calculation counts an undetected species as nothing, as EPA’s own per-home data does. That places the same home about 3.5 points higher. When the report gives the national median, it says which convention the median is on.

The ERMI’s developers put the score’s own standard deviation at about 3 (Section 4).

2.2 The Group

ExplainmyERMI organizes ERMI panel results into four Groups: Low, Average, Elevated and High Mold Load. These four Groups are not the panel’s Group 1 and Group 2 species lists. The Group is a fixed, deterministic reading of the pattern of elevation across the panel against the national reference values.

Its design is informed by two sources. The first is indoor mycology, in particular the ecological distinction between the water-damage indicator species and the species typical of outdoor air and ordinary indoor environments. The second is the multiplier-based reading of ERMI panels used in practice. The specific signals the Group reads and the way they are combined are proprietary and are not described on this site.

Three things about the Group can be stated. They come from our own analysis, and Section 8 says what a reader can check.

• It is not the ERMI score re-binned. On 1,790 public reference samples the Group and the ERMI quartile agree for about half of homes. When they differ, the reason is usually the subtraction described in Section 2.1. The score can come out low when both sets of species are elevated, and high when the common-indoor species are nearly absent. The Group reads those two kinds of home differently. Section 3 shows one example.
• More elevation never lowers it. Take two samples. If the first is at least as elevated as the second on every species, the first never gets a lower Group than the second. We checked every pair of reference homes, and many more possible patterns beyond them, and found no exception.
• A single very high species can sit in a low Group. A sample with one species far above its reference and little else raised can be Group 1 or Group 2. On the reference homes, 21 Group 1 homes carry an indicator species at 100 times reference or more. The report flags any such species on its own card in the individual mold results. The report’s Group 1 and Group 2 texts both say that one species can stand well above the rest in these Groups.

2.3 The Moisture Ladder

The ERMI panel measures 36 species selected by EPA researchers (Vesper et al. 2007). Each has a characteristic minimum water activity (A_w) at which it can grow. For most of the species that minimum is documented in the indoor-mycology literature. The report sorts the species into three bands by that minimum:

The three bands follow the primary, secondary and tertiary colonizer succession described in the building-mycology literature (Grant, Hunter, Flannigan and Bravery 1989; Adan 1994). Primary colonizers are xerophiles. They appear first, as conditions become favorable. Secondary colonizers are mesophiles. They follow when moisture persists. Tertiary colonizers are hydrophiles. They appear only on materials that have been wet. The German Federal Environment Agency’s mould guideline reproduces the framework with the thresholds we use (UBA 2017; English edition 2019, updated 2024).

The band a species falls into is a clue about the kind of moisture involved, not only about overall mold burden. A pattern led by xerophiles is generally associated with long-term high humidity. A pattern led by hydrophiles is generally associated with materials that have been wet. A balanced pattern is generally associated with more than one kind of moisture over time. These are associations from the published literature, not conclusions about a particular building.

The minimum water-activity values come from the published literature: Pitt and Hocking, Grant et al., Magan and Lacey, Segers et al. and others. The sources are listed with each species on its species page. For about ten species no species-level measurement exists, and we use a value estimated from closely related species or from the species’ colonizer class.

For two species we use the higher value reported on building materials in place of the laboratory minimum, because customers are reading buildings, not culture plates. For Stachybotrys chartarum that value comes from the UBA guideline. For Rhizopus stolonifer it comes from the range reported for its fungal order.

What the Ladder reports. The report draws the three bands as the Moisture Ladder. All 36 species are ranked on it by the moisture each needs, and the customer’s results are plotted on it. From the elevated species the report names a leading band. It says whether the pattern is concentrated in that band, leans toward it, or is spread across bands. When too little is elevated to support a reading, it says the pattern is too thin to read and names no band. On the reference homes that is the reading for 88% of Group 1 homes and 37% of Group 2 homes, and for almost none in Groups 3 and 4.

The rule by which the Ladder weighs the bands is part of the proprietary framework. The Ladder reads the dust, not the building. It says which kind of moisture the elevated species are associated with. It does not say whether, where or when that moisture occurred.

2.4 HERTSMI-2

The report also shows the HERTSMI-2 score, a five-species score developed by Dr. Ritchie Shoemaker, calculated with his published point thresholds. The “safe”, “caution” and “unsafe” ranges belong to that published method. They are not our determination about any building, and we are not affiliated with Dr. Shoemaker.

The five HERTSMI-2 species are among the 26 indicator species on the panel. On the reference homes, HERTSMI-2 and the Group often point the same way.

2.5 The building-context sections

Beyond the Group, each report uses two more kinds of information. The first is the building details the customer enters on the order form: year built, building type, foundation, heating and cooling, ventilation, known water events and optional indoor humidity readings. The second is outdoor climate data for the sample’s ZIP code. Those inputs shape four parts of the report:

• Connecting the dots, within Your Results: a plain-language account of the moisture conditions commonly associated with this species pattern in buildings of this type and climate.
• Walk-through Guide: a printable checklist of places where moisture problems commonly occur in this type of building. Every item is something to look at, not something to fix.
• Suggested Areas to Explore: topics a customer may want to raise with a qualified professional.
• Climate Context: a year of outdoor dew point for the location and, in colder climates, an estimate of when wall cavities are prone to condensation. The estimate assumes typical wall insulation for the building’s age. It is not a measurement of the customer’s walls.

These sections are general educational information. They describe conditions commonly associated with a species pattern in buildings of that type. They are not findings about the customer’s property. We do not visit, inspect, sample or survey any building. For the building itself, the sections rely only on what the customer tells us, and they cannot account for anything we were not told. They have not been checked against inspection findings at any property.

The sections are rule-based and deterministic. They use no machine learning, no hidden parameters and no per-customer model fitting. The templates and modifier rules draw on the published indoor-mycology and residential building-science literature:

• water-activity minima per species (Grant et al. 1989; Pitt and Hocking 2009; WHO 2009);
• ASHRAE Standards 62.1, 62.2, 55 and 160;
• the heat-and-moisture transport principles described by Lstiburek and Carmody and by Harriman and Lstiburek;
• the ASHRAE Position Document on Limiting Indoor Mold and Dampness in Buildings (2021, reaffirmed 2024);
• the time-of-wetness framework of Adan (1994) as operationalized by Sedlbauer (2001);
• indoor-aerosol research on fragment-bound exposure (Górny et al. 2002; Cho et al. 2005).

These sections cannot be tested against outcomes today. That would take building details, ERMI results and inspection findings for the same homes, and no public data set has them. In the report, the building-context text ends by saying that the mechanisms it lists are prior expectations for this kind of building, not findings about your specific building.

3. Two homes with the same score

Two real homes from the public reference data show what the Group adds to the score. Both have the same ERMI score. Home identifiers are EPA’s sample IDs.

The scores match because of the subtraction. In the first home, four elevated common-indoor species are subtracted from three elevated indicator species. The second home has one elevated species and almost nothing to subtract. The Group reads the first home as broadly elevated and the second as a single species standing alone. Neither reading says what is happening in either house. Both say what is in the dust.

The same pattern recurs across the reference data. Among the 236 homes scoring between 4 and 6, 31% are Group 1, 45% Group 2, 21% Group 3 and 3% Group 4. The Ladder’s leading band is humidity for 44% of them, condensation for 16% and wet for 3%. For the other 37% the pattern is too thin to read.

4. How much a single result can move

Any interpretation of an ERMI result inherits the uncertainty in the 36 numbers. The ERMI’s developers state that each of the 36 assays has a standard deviation of about half a log, so that any ERMI value has a standard deviation of about 3. In their words, “a home with an ERMI value of 2 is not different from a 1 or a 4” (Vesper et al. 2007). In the same study, 435 samples re-analyzed at EPA’s own laboratory correlated 0.73 with the commercial laboratories’ results for the same samples. Repeat samples from the same home months apart agree less well still (Spearman 0.71 over 1 to 7 months, Kamal et al. 2014).

We simulated noise of that size on the 1,790 reference samples. Each detected species value was multiplied by a random factor, drawn independently for each species from a log-normal distribution with a standard deviation of half a log. This was done 200 times per home, and each label was recomputed. Undetected species were left undetected, which understates real variation.

At the developers’ stated assay precision, about 31 homes in 100 would land in a different Group on re-analysis. About 29 would land in a different ERMI quartile, and about 20 in a different HERTSMI-2 category. The Group is no steadier than the score. These simulated figures are best read as a floor. The published disagreement between laboratories is larger than noise of this size alone would produce.

Two more results point the same way. A laboratory that reads every species twice as high moves the Group more often than the score, because the score’s subtraction cancels a shared shift. A single species reported twice as high or twice as low would change the Group of about four homes in ten, against about one in eight for the ERMI quartile.

Among reference homes sampled again thirteen years later, a quarter kept their Group and a little over a third kept their ERMI quartile.

So read the Group as a band, not a verdict.

5. Limitations

ExplainmyERMI’s interpretation is bounded by the limitations of the underlying ERMI test and of our own analysis. They are listed here:

• The Group has not been tested against outcomes. The public survey data holds dust results and the year each home was built. It holds nothing about occupant health, inspection findings, water damage or remediation. Nothing on this page shows that a Group corresponds to illness, to what an inspector would find, or to whether a home is safe.
• The figures on this page are our own. They come from our own scripts, and nobody outside the company has repeated them. Section 8 says what a reader can check.
• Single-sample variance. Repeat sampling from the same home can give a different score. Every report says so in its disclaimer and again in its FAQ. In one published cohort (Kamal et al. 2014), 52 paired samples taken 1 to 7 months apart had mean ERMI values of 12.8 and 15.2. Repeat scores correlated with the first ones (0.71), and individual homes moved in both directions. ERMI was designed for cohort-level research, not individual-home diagnosis. We treat any given dust result as a signal integrated over weeks to months, not a point-in-time finding.
• Inter-laboratory variance is an upstream limitation we cannot correct. EPA’s Office of Inspector General (Report 13-P-0356, August 2013) noted that two MSQPCR-licensed laboratories testing the same home may produce different results, and the published figures bear that out (Section 4). ExplainmyERMI’s interpretation is anchored to the species values the customer’s chosen laboratory reported. We cannot detect or correct for differences that arise before the data reaches us.
• The reference values date from the mid-2000s. The geometric means behind every elevation factor come from the first American Healthy Homes Survey. A second survey in 2018 and 2019 found broadly similar values for the 36 species, with somewhat higher average ERMI scores (Vesper et al. 2021). Any single home is being compared with a national average that is now some years old. If typical readings have risen since, a home sampled today will read higher against the old reference than it would against a current one.
• The sample may not have been collected the way the reference samples were. In the surveys, dust was vacuumed from a set area of floor in the living room and a bedroom (Vesper et al. 2007). Many consumer samples are collected differently: with a cloth, from other surfaces, from one room, or from dust of a different age. An elevation factor compares the customer’s value with the reference as if both were collected the same way. The further the collection departs from the survey method, the weaker that comparison is.
• One national reference is used for every region and season. The reference values are not adjusted for region, climate or season. Outdoor mold levels differ between climates and through the year, and outdoor mold is one source of the mold in house dust. A reading that is ordinary for one climate may count as elevated against the national reference, and the reverse.
• No source localization. Dust testing measures the species pattern at the sampled location but does not tell you where in the building the contamination originates. A wet-band pattern in whole-house dust could come from a single wall cavity, a crawlspace corner, or a leak in a remote utility room. On-site inspection by a qualified, independent professional is the appropriate way to localize sources.
• No current-vs-historical distinction. Dust integrates over weeks to months. A species pattern may reflect current conditions. It may reflect recent past conditions, for example before a dehumidifier was installed. Or it may reflect material that hosted growth in the past and now hosts residual colonization. The report flags these alternative readings when the customer’s building details suggest a mismatch is plausible.
• No label from a single sample is a settled property of a home. Section 4 gives the numbers.
• EPA does not recommend ERMI outside research. Section 6 gives EPA’s position in EPA’s words.

6. EPA’s position on ERMI

EPA, whose researchers developed ERMI, recommends against its use outside research. EPA’s public mold-information page is at epa.gov/mold/should-i-test-or-sample-mold-my-home-using-environmental-relative-moldiness-index-or-ermi. It was last updated on 8 September 2026 and was read for this page on 10 October 2026. The page asks: “Should I test or sample for mold in my home using the Environmental Relative Moldiness Index, or ERMI?” It answers: “No. The Environmental Relative Moldiness Index, or ERMI, developed by U.S. Environmental Protection Agency researchers, is a research tool and is not recommended for use except as a research tool.”

EPA gives its reason in its own words. Its research page says ERMI “has been peer reviewed for research purposes but has not been validated for non-research purposes.” Its fact sheet continues: “For this reason, EPA does not recommend the routine public use of ERMI in homes, schools, or other buildings,” and adds that because the approach “has not been validated through a multi-lab study, we cannot comment on the accuracy of information obtained from others using ERMI.” This position was set out in the EPA Office of Inspector General Report 13-P-0356 (August 2013). It has been restated in EPA fact sheets (2014, 2016 and September 2021) and on EPA’s public mold pages.

Our position is narrow. We do not perform, sell or recommend the ERMI test. We do not claim EPA endorsement of the test, of our interpretation of it, or of this page. EPA’s recommendation is reproduced here so that customers can read it directly. The decision to obtain an ERMI test from a laboratory of their choice rests with the customer. Once that result exists, we provide an educational interpretation of it.

7. What ExplainmyERMI is not

To make the scope of the service explicit:

• Not a mold inspection or assessment. We do not visit properties, collect samples, or provide remediation guidance. We are not a licensed mold assessor in Texas or any other jurisdiction.
• Not medical advice. We do not diagnose conditions, prescribe treatments, or recommend medical evaluation specific to a customer’s health.
• Not a substitute for on-site assessment. The report is a starting point. Its Walk-through Guide lists things a customer can look at. It does not tell them what is wrong or what to do about it. Confirming whether a moisture or mold problem exists, and where it comes from, takes an on-site inspection by a qualified, independent professional.
• Not regulatory compliance documentation. ERMI is not an approved test for any regulatory mold-assessment purpose in any U.S. jurisdiction we are aware of. Customers who need regulatory documentation should consult a licensed mold assessor in their state.
• Not affiliated with EPA or any government agency. EPA researchers developed the ERMI research tool, and EPA has licensed the underlying MSQPCR technology to private laboratories. EPA has not endorsed any interpretation service, including ours.

8. What you can check

The data is public. EPA publishes the AHHS II species values and per-home summaries for both surveys (U.S. EPA Office of Research and Development, doi:10.23719/1519263). From that data and the published formulas, a reader can reproduce the parts of this page that do not depend on our Group:

• the ERMI quartiles under either non-detect convention;
• how often the ERMI quartile and the HERTSMI-2 category change under the noise described in Section 4;
• how many re-sampled homes kept their ERMI quartile between the two surveys.

The Group’s rules are proprietary and are not published, so the Group figures cannot be re-derived from this page. We say so rather than claim otherwise. In-house, the analyses were run with seeded scripts that call the production classifier directly, and their full output is kept.

References

The primary sources this page draws on.

1. Adan, O. C. G. (1994). On the fungal defacement of interior finishes. Doctoral thesis, Eindhoven University of Technology. DOI: 10.6100/IR427806.
2. ASHRAE. Position Document on Limiting Indoor Mold and Dampness in Buildings (approved 2021, reaffirmed 2024); ANSI/ASHRAE Standards 62.1-2022, 62.2-2022, 55 (2017 edition) and 160-2021.
3. Brasel, T. L., Douglas, D. R., Wilson, S. C., & Straus, D. C. (2005). Detection of airborne Stachybotrys chartarum macrocyclic trichothecene mycotoxins on particulates smaller than conidia. Applied and Environmental Microbiology 71(1): 114–122. DOI: 10.1128/AEM.71.1.114-122.2005.
4. Cho, S.-H., Seo, S.-C., Schmechel, D., Grinshpun, S. A., & Reponen, T. (2005). Aerodynamic characteristics and respiratory deposition of fungal fragments. Atmospheric Environment 39(30): 5454–5465. DOI: 10.1016/j.atmosenv.2005.05.042.
5. EPA Office of Inspector General. (August 22, 2013). Public May Be Making Indoor Mold Cleanup Decisions Based on EPA Tool Developed Only for Research Applications (Report 13-P-0356). U.S. Environmental Protection Agency. EPA OIG Report page.
6. U.S. Environmental Protection Agency. Should I test or sample for mold in my home using the Environmental Relative Moldiness Index, or ERMI? epa.gov/mold (page last updated 8 September 2026). EPA page.
7. U.S. Environmental Protection Agency. Environmental Relative Moldiness Index (ERMI) research page (last updated 29 September 2026) and the September 2021 fact sheet, Center for Environmental Measurement and Modeling. EPA research page.
8. U.S. EPA Office of Research and Development. AHHS 1 and 2 SciHUB 8-17-20 (data set). DOI: 10.23719/1519263.
9. German Federal Environment Agency (UBA). Guideline for the prevention, detection and remediation of mould in buildings. 2017; English edition 2019, updated 2024. Indoor Air Hygiene Commission. ISSN 2363-8311. English PDF.
10. Górny, R. L., Reponen, T., Willeke, K., Schmechel, D., Robine, E., Boissier, M., & Grinshpun, S. A. (2002). Fungal fragments as indoor air biocontaminants. Applied and Environmental Microbiology 68(7): 3522–3531. DOI: 10.1128/AEM.68.7.3522-3531.2002.
11. Grant, C., Hunter, C. A., Flannigan, B., & Bravery, A. F. (1989). The moisture requirements of moulds isolated from domestic dwellings. International Biodeterioration 25: 259–284. DOI: 10.1016/0265-3036(89)90002-X.
12. Harriman, L., & Lstiburek, J. (2009). The ASHRAE Guide for Buildings in Hot and Humid Climates (2nd ed.). ASHRAE.
13. Kamal, A., et al. (2014). Applicability of the Environmental Relative Moldiness Index for quantification of residential mold contamination in an air pollution health effects study. Journal of Environmental and Public Health 2014: 261357. DOI: 10.1155/2014/261357.
14. Lstiburek, J., & Carmody, J. (1993). Moisture Control Handbook: Principles and Practices for Residential and Small Commercial Buildings. Van Nostrand Reinhold.
15. Nevalainen, A., Pasanen, A.-L., Niininen, M., Reponen, T., Kalliokoski, P., & Jantunen, M. J. (1991). The indoor air quality in Finnish homes with mold problems. Environment International 17: 299–302. DOI: 10.1016/0160-4120(91)90015-I.
16. Pitt, J. I., & Hocking, A. D. (2009). Fungi and Food Spoilage (3rd ed.). Springer. DOI: 10.1007/978-0-387-92207-2.
17. Sedlbauer, K. (2001). Vorhersage von Schimmelpilzbildung auf und in Bauteilen (Prediction of mould fungus formation on the surface of and inside building components). Doctoral dissertation, Universität Stuttgart.
18. Strachan, D. P., Flannigan, B., McCabe, E. M., & McGarry, F. (1990). Quantification of airborne moulds in the homes of children with and without wheeze. Thorax 45(5): 382–387. DOI: 10.1136/thx.45.5.382.
19. Thatcher, T. L., & Layton, D. W. (1995). Deposition, resuspension, and penetration of particles within a residence. Atmospheric Environment 29(13): 1487–1497. DOI: 10.1016/1352-2310(95)00016-R.
20. Vesper, S., McKinstry, C., Haugland, R., Wymer, L., Bradham, K., Ashley, P., Cox, D., Dewalt, G., & Friedman, W. (2007). Development of an Environmental Relative Moldiness Index for US homes. Journal of Occupational and Environmental Medicine 49(8): 829–833. DOI: 10.1097/JOM.0b013e3181255e98.
21. Vesper, S., Wymer, L., Cox, D., Dewalt, G., Pinzer, E., Friedman, W., & Ashley, P. J. (2021). The Environmental Relative Moldiness Index reveals changes in mold contamination in United States homes over time. Journal of Occupational and Environmental Hygiene 18(1): 35–41. DOI: 10.1080/15459624.2020.1844892.
22. WHO Regional Office for Europe. (2009). WHO Guidelines for Indoor Air Quality: Dampness and Mould. World Health Organization. WHO publication page.