Science, technology and consultancy services

Cheese quality

Causes of early gas production in Cheddar cheese

About this article

blown cheeseThis article is based on a paper, summarising many years of research, published in the International Journal of Dairy Technology. The paper “Mullan, W.M.A. (2000). Causes and control of early gas production in Cheddar cheese. International Journal of Dairy Technology. 53, 63-68” can be downloaded from the Wiley Interscience website. Since then the author has continued to work in this area and in particular with major Cheddar cheese manufacturers in the US. This newer work is not discussed in this article.

Currently (in 2012) many cheesemaking plants are experiencing open texture problems including unwanted slits/cracks in cheese due to unwanted gas production. Some of these problems are caused by the growth of thermophilic / thermoduric lactic acid bacteria in biofilms in pasteurizers. Normal caustic cleaning will not eliminate these and enhanced cleaning and sanitation procedures are required.  While this contribution may not specifically deal with these problems this area can be discussed further in the forums.




Early gas production in Cheddar cheese is a sporadic and recurrent problem. In this article the major causal agents of
unwanted gas production are discussed and potential gas producing organisms in Cheddar cheese are identified. Many of these agents will also cause problems in other cheese varieties. Early gas production in Cheddar cheese results from a number of interacting factors including lactose and citrate levels in the curd, the temperature of curd/cheese during pressing and curing, the salt in moisture level in the cheese and the levels of gas-producing, non-starter lactic acid bacteria in the cheese. Note added in June 2010: Some of these problems can be caused by the growth of thermophilic / thermoduric lactic acid bacteria in biofilms in pasteurizers and their growth to high population densities during cheese making and pressing.

Early gas production in 18-kg blocks of commercial Cheddar cheese is a well known, if not fully understood, phenomenon1. Incidents of gas production tend to be sporadic and recurrent and have probably been experienced at most cheese-making plants. Below the cheese on the left displays a "loose" bag due to gas production. The extent of gas production can be such that the 'slats' can buckle or even break due to the pressure generated.

Cheese on the left displays a loose barrier bag caused by gas production

Open texture, slits and fissures in cheese producing excessive gasAlthough some incidents of unwanted gas production can be explained by poor cheese-making practices, including bad hygiene, or starter failure, this is not universally true. The author has investigated incidents of early gas production in commercial cheese plants (within 3-8 weeks of manufacture) over many years in which the cheese graded normally, and was within acceptable limits for salt, moisture and pH. The cheeses studied were also free from significant levels of gas-producing non-lactic acid bacteria.Note open texture including unwanted slits/cracks in cheese can be caused by unwanted gas production. Note the cheese on the left displays cracks and fissures due to unwanted gas production. Another image showing more extensive slits/ fissures is also shown.


Products of the lactate fermentation, combined with control of the moisture and salt levels in the final cheese, good hygiene, and the use of good quality pasteurised milk effectively limit the range of bacteria which can produce gas in Cheddar cheese 6. Although gas can be produced from a wide range of compounds present in cheese, lactose, lactate, citrate and urea are the major substrates involved (Table 1).Note other substrates may be more important in the open texture problems created due to growth of thermophilic / thermoduric lactic acid bacteria including Streptococcus thermophilus -like organisms in pasteurizers during long production runs.

The involvement of Clostridium tyrobutyricum in gas production in brined-cheeses such as Gouda and Grana Padano has been well established. However, this clostridium would not normally be expected to cause problems in Cheddar cheese of satisfactory salt, acid and moisture content7. Some


Click for help with cracks or slits in cheese



homofermentative bacteria e.g. Lactobacillus casei, and Lactococcus lactis ssp. lactis biovar. diacetylactis can produce CO2 from citrate and have been implicated in the blowing of film-wrapped cheese8. The potential for gas production by heterofermentative lactobacilli has been known since the work of Sherwood9. Coliforms are usually only a problem when the starter fails, due to bacteriophage infection or antibiotic residues in milk. Under certain conditions, urease producing strains of Streptococcus thermophilus can produce gas in Cheddar cheese 10. Since Streptococcus thermophilus can grow in the regeneration section of pasteurisers relatively high levels of this thermophile may occasionally occur in pasteurised milk. Note it may be more accurate to describe the Streptococcus thermophilus isolates from pasteurizers as Streptococcus thermophilus- like since they frequently differ in a number of important respects from normal dairy strains e.g. many are much more tolerant to salt (NaCl).  Incidentially the galactose produced by Streptococcus thermophilus can be used as a substrate for gas production by non starter lactic acid bacteria. Note Sherwood9 first established that the addition of leuconostocs to milk for Cheddar cheese manufacture gave rise to gas and open texture in the cheese. Because of their sensitivity to salt and high acidity propionibacteria would not be expected to produce gas in Cheddar cheese. Examination of the gas produced by the major gas-producing organisms (Table 1) indicates that the identification of the gas present in blown cheese may have value in identifying the gas producing species.


Determination of the causal agent(s) of gas production in cheese can be difficult. The numbers of the micro-organism concerned may have declined to a low, apparently insignificant level at sampling. This can be overcome by serial sampling during maturation. Some micro-organisms are not distributed uniformly in cheese, for example lactobacilli may be found in fissures or curd junctions. Sampling schemes should be designed to take account of the potential for non-random distribution of the gas producing agent

The selective agar media used may be inadequate; e.g., some leuconostocs and pediococci will form colonies on Rogosa agar originally developed for work with lactobacilli3. Certain media may allow the growth of starter lactococci.

Microscopic studies of 'gassy' cheese can also yield inconclusive results. Leuconostocs, for example, may appear as small rods, cocco-bacilli or cocci. Heterofermentative lactobacilli may also appear as cocco-bacilli

Where a particular micro-organism is suspected of being the causal agent, current molecular biology techniques involving polymerase chain reaction (PCR) amplification of DNA sequences combined with species-specific DNA probes can be used to confirm the identify of the micro-organism. Klijn et al.12 have used PCR and a species specific-DNA probe to confirm Cl. tyrobutyricum as the causative agent of late blowing in experimental cheese.


The results of initial microbial studies (Table 2) at one plant (factory X) revealed that with the exception of catalase-negative citrate-utilisers, 'total count' on milk agar and 'counts' on Rogosa agar, microbial levels in blown and 'normal' cheeses were similar. Because bacterial numbers on Rogosa agar were relatively low <1 106 cfu/g of cheese, these bacteria were not studied further at this stage but in later studies, blown cheeses were found to have 10-100 fold higher counts on Rogosa agar compared with normal cheese at factory X.

The scheme described by Billie et al.3 was used to identify presumptive lactic acid bacteria to the genus level and the API-CHL computer-assisted identification system to classify isolates to the species level. Detailed investigation of some of the catalase-negative citrate-fermenting isolates revealed that they resembled Lc. lactis ssp. lactis biovar. diacetylactis in their general properties. The data obtained were consistent with the Lc. lactis ssp. lactis biovar. diacetylactis isolates originating from the mixed strain starter cultures used.

These observations prompted us to study citrate levels in milk to determine if there was a relationship between citrate concentration and the incidence of gas production in factory X (Fig. 1). Although lowest citrate levels occurred during April, May and June, the incidence of gas production in this period was not markedly different than at other times during the period of study.

Pilot-scale cheese-making experiments were used to define the conditions required for gas production by Lc. lactis ssp. lactis biovar. diacetylactis in experimental cheese of satisfactory chemical composition. Sufficient gas to give overt "blowing" of barrier bags was only produced when mixed-strain cultures containing high levels of Lc. lactis ssp. lactis biovar. diacetylactis were used and the citrate level in cheese ex-press was 0.07% (w/w) or higher. To obtain this level of citrate, the cheese milk always had to be supplemented with additional citrate.

Attempts were made to correlate citrate levels in cheese, ex-press, with the development of gas in cheese at factory X. Some of the data obtained are shown in Table 3. These results show that gas production occurred in cheeses containing low and high levels of residual citrate and suggest that early gas production at factory X was not simply due to the activity of Lc. lactis biovar. diacetylactis. These findings also indicated that gas was being produced from substrates other than citrate.

During this study an opportunity arose to investigate the "blowing" of 4-kg wheels of Leicester cheese at another factory. The Leicester wheels had "blown" after 3 weeks storage at 7ºC.

Blowing of 4-kg wheel of Leicester cheese

Analyses revealed that:

i. salt/moisture (S/M) levels were satisfactory (ca. 5%)
ii. pH values were acceptable (ca. 5.1)
iii. levels of gas-producing non-lactic acid bacteria and yeasts were within acceptable limits.
iv. cheese samples gave counts >l08 cfu/g on Rogosa agar
v. the major gas present was mainly carbon dioxide with small concentrations of nitrogen and oxygen.

After discussions with production staff, it became apparent that a new starter had been introduced into the starter rotation sequence and that it was this culture which had been used to make the defective cheese.

A can of unopened starter of the same code and batch number as that used to make the bulk starter used in the manufacture of the blown cheeses was shown to contain organisms capable of inducing high levels of gas in two experiments:

i. when reconstituted skim milk (RSM) was inoculated with culture, with or without added citrate, and the contents sealed with wax and agar, gas sufficient to displace the seals in control and citrated- milk was produced; and
ii. gas was produced when the culture was used as a starter in the manufacture of experimental Cheddar cheese.

NOTE! the use of simple milk based media as in i above cannot be relied upon to confirm the gas producing potential of isolates in cheese.

Detailed examination of the starter revealed the presence of high levels of strains (>50%) that grew on Rogosa agar. Some characteristics of the Rogosa isolates are given in Table 4. The relative salt insensitivity of the isolates and the fact that they grew at 6°C were of interest. The characteristics of these isolates were consistent with those of Leuconostoc mesenteroides ssp. dextranicum.
In controlled cheese-making experiments, gas production was induced using a 0.00l% inoculum of Leucon. mesenteroides ssp. dextranicum (isolate HM8/10). A section of experimental cheese showing the open texture of cheese produced using HM8/10 is shown on the left.

experimental cheese exhibiting slits due to gas productionIn view of these results, 13 of the mixed strain starters at factory X were screened for the presence of organisms capable of growing on Rogosa agar. Seven of the starters contained strains which grew on Rogosa agar. Unlike the situation described with the starter used at factory Y, starters at factory X generally contained <0.0000l% i.e. 10 2-103 cfu/ml of heterofermentative strains. Most Rogosa agar-isolates were capable of growth in 6% (w/w) NaCl and at 6°C.

Further study of cheese from factory X generally revealed that counts on Rogosa agar for "gassy" cheese differed from normal cheese by a factor of 10-100. Levels of non-starter lactic acid bacteria (NSLAB) generally increased with age of cheese. The Rogosa agar isolates had properties consistent with leuconostocs or heterofermentative lactobacilli. Most were also salt tolerant and capable of growth at 6°C.

An examination of water, rennet, pasteurised whole milk and annato at several plants for NSLAB (detection limit 1 cfu/ml) was unsuccessful. Low levels of NSLAB generally <100 cfu/ml were found in raw milk. Variable levels <10-103 cfu/ml were found in tanker skim-milk used to standardise milk for cheese manufacture. Characterisation of these bacteria revealed properties consistent with L. fermenti. These isolates were also capable of growth at 6ºC and at S/M levels of 5.5%.


Under conventional Cheddar cheesemaking conditions, using only lactococcal cultures, the four main substrates used by microorganisms for gas production are lactate, lactose, citrate and urea. Note gas can be formed from other substrates and these may be involved in particular incidents; e.g. when cultures containing Str. thermophilus are used.

In the incidents of early gas production studied by the author and summarised here no evidence was found to suggest that lactate or urea was involved in gas production. The volume of gas potentially available from lactose and citrate is shown in Table 5. Note that CO2 is very soluble in water and that 'normal' cheese can contain significant concentrations of this gas. Clearly the volume of gas potentially available from low levels of these substrates is substantial.

There are a number of strategies that can be used to limit undesirable gas production in cheese. These include:

Choice of starter. Leuconostocs and Lc. lactis ssp. lactis biovar. diacetylactis can produce gas from lactose and citrate, and citrate, respectively. One method of control is to use cultures containing only strains of Lc. lactis ssp. cremoris and/or lactis. However, using only lactococci which do not produce gas from citrate may not always solve gas production problems. Some homofermentative lactobacilli e.g. L. casei and L. plantarum can produce gas from citrate. Due to the presence of high concentrations of citrate in cheese made with Lc. lactis ssp. cremoris and/or Lc. lactis ssp lactis, gas may be produced if citrate-utilising NSLAB such as L. casei or L. plantarum reach high levels. Since NSLAB may constitute a significant part of the indigenous microbial flora of a factory, it may take some time to to reduce them to low levels.

Apart from controlling gas production from citrate-utilising NSLAB by good hygiene, there is also the option of producing curd of low citrate concentration. Curd of low citrate-content can be produced in a number of ways including the controlled use of starters containing citrate-utilisers.

Cheese produced using starters containing Streptococcus thermophilus has the potential to gas if there is significant residual galactose in the cheese. This possibility combined with the urease activity of most strains needs to be considered carefully when plants are experiencing gas production problems including 'slitty' cheese.

Cheese-making conditions. To ensure that lactose levels in cheese after packing are low and that it is metabolised rapidly during the first 24 hours of curing, the acidity of curd at salting should be as high as possible. S/M levels should be between 4.5-5.5% since marked inhibition of lactose utilisation occurs at S/M levels >5.8%. Phage levels in curd at pressing must be controlled to prevent starter cell lysis that could result in 'sweet cheese' containing high concentrations of residual lactose6.

There is also an option of producing curd of low citrate concentration as discussed previously.

Curing temperatures and pressing systems. The length of time required for the temperature at the centre of cheese blocks to reach <10°C is influenced by type of pressing system used, whether pre-cooling was used and the method of handling and storing the packaged cheese. Results indicated that, in some cheese plants it took some 3-6 days from the start of pressing for the temperature at the centre of cheese blocks to reach <10°C. These conditions are ideal for the rapid growth of NSLAB.

Cheddar cheese should be cooled rapidly to <10°C and reach a core temperature of 7°C as soon as possible. Consideration should be given to pre-chilling cheese to 7ºC, using a blast-chilling system, before placing cheese in ripening rooms. Low temperatures will retard the growth of NSLAB, while the starter lactococci utilise the residual lactose thus reducing the potential for gas production from lactose.

Barrier bags permeable to carbon dioxide. Barrier bags of increased permeability to CO2, but of only slightly increased permeability to O2, can be purchased and may be useful in some instances. Because of their greater permeability to oxygen, the potential for mould growth on the packaged cheese is increased. They are also more expensive than conventional bags, and because of the cost difference, it may be cheaper to repack blown cheese.

Biological control . Some forms of gas production, e.g. that caused by clostridia, can be controlled by the use of nisin-producing starter bacteria.

Hygiene. It is important to minimise the numbers of gas producing NSLAB in cheese. Good cleaning and sterilising procedures are required. There may also be a requirement for aerosol disinfection and/or fumigation of the cheese-making area.

Pasteuriser operation. Coliform contamination can sometimes be linked with pasteurisation faults including incorrect pasteuriser operation. Mixing of pasteurised and non-pasteurised milk due to holes in plates or leaky gaskets occurs occasionally and should be considered in studies of "gassy" cheese. In situations where long pasteurisation runs occur, there may be merit in testing milk from the regeneration section of pasteurisers for the presence of Streptococcus thermophilus.


Note the information presented here is in summary form for commercial reasons and additional information will generally be required to investigate many incidents of gas production including those causing cracks/slits in the cheese.

Starter testing. If it is decided to eliminate starters, that produce gas from citrate, then those that produce gas in citrated-milk should not be used. Starters containing high levels of heterofermentative lactic acid bacteria capable of growth in cheese can be detected based on agar seal displacement using Rogosa broth or RSM supplemented with yeast extract. More sophisticated testing methods will be required when the starter contains only low concentrations of potentially fault-causing microorganisms.

pH and salt in moisture levels. pH and S/M values in 24-h old Cheddar cheese produced from pasteurised milk should be within the range 4.9 -5.2 and 4 - 5.8% (w/w), respectively.

Counts on Rogosa agar. Counts on this medium should be low in 1-week old cheese. The author has observed NSLAB counts on this medium of <1x105/g after 5 months storage for cheeses from some plants. Note while this medium can be useful modifications may be required to aid the successful investigation of some incidents of unwanted gas production.


While it has been possible to ascribe some incidents of early gas production in Cheddar cheese to the activity of citrate-utilising and/or heterofermentative strains in mixed strain cultures, we have not been able to establish a definitive causal relationship between starter type and early gas production in Cheddar cheese.

Many of the commercial mixed-strain cultures studied over a number of years contained low levels of heterofermentative gas-producing, non-starter lactic acid bacteria capable of growth at 6
ºC and at high S/M levels, conditions found during the maturation of cheese. Evidence for the involvement of heterofermentative, psychrotrophic, salt-insensitive organisms capable of growth on Rogosa agar in early gas production was found.

Heterofermentative lactic acid bacteria have been isolated from raw milk and tanker samples of pasteurised skim. It is probable that the gas-producing NSLAB in cheese are derived from contamination of the factory environment by bacteria present in raw milk, tanker skim-milk and/or mixed-strain cultures. If it can be shown that small numbers of gas-producing NSLAB survive pasteurisation, then there may be a direct raw milk quality or a tanker skim-milk quality dimension in early gas production in cheese.

Contamination of cheese milk with low levels of gas-producing lactic acid bacteria may not always result in blown cheese.

Gas production in cheese of normal chemical composition and pH results from a number of interacting factors, including the starter used, lactose and citrate levels in the curd, temperatures of curd/cheese during pressing and curing, the salt sensitivity of the starter, the S/M level in the cheese, the levels of gas-producing NSLAB bacteria in the cheese, and the level of phage-induced cell lysis in the curd at pressing and during early cheese maturation.

1 Prentice G A and Brown J V (l983) The microbiology of Cheddar cheese manufacture. Dairy Industries International 48(7) 23-26.
2 Billie P G, Mullan W M A and Espie W E (l985) Observations on the effectiveness of agar media for the enumeration of leuconostocs from cheese and starter cultures. Irish Journal of Food Science and Technology 9 78. (Abstract).
3 Billie P G, Espie W E and Mullan W M A (1992) Evaluation of media for the isolation of leuconostocs from fermented products. Milchwissenschaft 47 637-640.
4 Mullan W M A, Barnes E D and Espie W E (1983) Studies on gas production in Cheddar cheese. Irish Journal of Food Science and Technology 8 85. (Abstract).
5 Mullan W M A, Espie W E and Barnes E D (1984) Role of commercial mixed strain starter cultures in incidents of gas production in Cheddar cheese. Annual Report on Research and Technical Work (1984) Department of Agriculture Northern Ireland p.321.
6 Mullan W M A (l986) Bacteriophage induced starter problems. Dairy Industries International 5l (11) 40-43.
7 Kleter G , Lammers W L and Vos E A ( 1984) The influence of pH and concentration of lactic acid and NaCl on the growth of Clostridium tyrobutyricum in whey and cheese. 2. Experiments in cheese. Netherlands Milk and Dairy Journal 38 31-41.
8 Fryer T F, Sharpe M E and Reiter B (1970) Utilisation of milk citrate by lactic acid bacteria and 'blowing' of film wrapped cheese. Journal Dairy Research 37 17- 29.
9 Sherwood I R (l939) The relation of certain lactic acid bacteria to open texture in Cheddar cheese. Journal Dairy Research 10 326-335.
10 Tinson W, Broome M C, Hillier A J and Jago G R (1982) Metabolism of Streptococcus thermophilus. Australian Journal Dairy Technology 37 14-16.
11 Fryer T F (l969) Microflora of Cheddar cheese and its influence in cheese flavour. Dairy Science Abstracts 31 471.
12 Klijn N, Nieuwenhof F F J, Hoolwerf J D, Waals, C B Van Der, Weerkamp, A H (1995) Identification of Clostridium tyrobutyricum the causal agent of late blowing in cheese by species-specific PCR amplification. Applied and Environmental Microbiology 61 2919-2924.
13 Reddy M S, Vedamuthu E R, Washam C J and Reinbold G W (1972) Agar medium for the differentiation and enumeration of lactic streptococci. Applied Microbiology 24 947-952.
14 Mullan W M A and Walker A L (1979) An agar medium and a simple streaking technique for the differentiation of the lactic streptococci. Dairy Industries 44(6) 13-17.
15 Rogosa M, Mitchell J A and Wiseman R F (1951) A selective medium for the isolation and enumeration of oral and fecal lactobacilli. Journal Bacteriology 62 132-133.

How to cite this article

Mullan, W.M.A. (2003). [On-line]. Available from: . Accessed: 24 April, 2014. Revised June 2008; July 2008; May 2012.


We use cookies to improve our website and your experience when using it. To find out more about the cookies we use, and how to manage them, see our Privacy and Cookie Policy.

By clicking continue you are agreeing to accept cookies

'Learn more about managing cookies'