Supplement Material for Investigation of the pollution level and affecting factors of formaldehyde in typical public places in Guangxi , China

College of Environment and Resources, Guangxi Normal University, Guilin, Guangxi 541004, PR China Key Laboratory of Ecology of Rare and Endangered Species and Environmental Protection, Guilin, Guangxi 541004, PR China Key Laboratory of Karst Ecology and Environment Change, Guangxi Department of Education, Guilin, Guangxi 541004, PR China Guangxi Key Laboratory of Environmental Pollution Control Theory and Technology, Guilin, Guangxi 541004, PR China


INTRODUCTION
As national economies continue to expand, increasing attention has been given to indoor air pollution.Formaldehyde, emitted from building materials, is regarded to be a main source of indoor air contaminations (Schlink et al., 2004;Kim et al., 2006;Missia et al., 2010;Tran et al., 2012;Shinohara et al., 2014;Wetzel, 2014), which can produce serious environmental issues for indoor air quality and pose a severe threat to the health of more than one-half of the world's population who spend more than 90 percent of their time indoors (Bonamano et al., 2016;Zhang et al., 2016).Formaldehyde released from building material is a long-term chronic process, which causes accumulation of formaldehyde and can cause indoor formaldehyde concentrations to exceed allowable limits (Yang et al., 2001).Repeated or prolonged exposure to high levels of formaldehyde concentration in the indoor environment can result in serious respiratory diseases, eye irritation, headache, asthma, as well as in degenerative, inflammatory and hyperplastic changes of the nasal mucosa (Lim et al., 2011).Long-term inhalation of formaldehyde in the indoor environment has also been identified as an important source of Sick Building Syndrome (SBS) (Horemans et al., 2010;Nakaoka et al., 2014;Mentese et al., 2015).Consequently, formaldehyde is an important indicator of indoor air quality (IAQ) and as such has attracted wide scale public attention (Raw et al., 2004;Liu et al., 2008;Allen et al., 2016).Health impacts due to formaldehyde exposure should be an important consideration and are worthy of study (Sumin et al., 2010;Hwang et al., 2011;Szulejko et al., 2016).
China has developed a large tourist industry in recent years (Zhang and Gao, 2016), which has resulted in an annual influx of large numbers of international tourists (Yang et al., 2010).According to the latest statistics from the China National Tourism Administration, more than 65 million foreign tourists visited China in the first half of 2016 and these numbers are expected to increase annually.Indoor air quality in various large scale indoor entertainment venues, such as shopping malls, department stores, supermarkets.has received much attention in the most of popular tourist locales (Tang et al., 2005;Chan et al., 2011;Gao et al., 2012;Tao et al., 2015;Fang et al., 2016;Shang et al., 2016).Consequently, air quality is a key item of interest in the improvement of indoor environments.Nevertheless, it has been reported that the highest formaldehyde concentration (0.355 mg m -3 ) has been measured in thirteen new renovated hotels in Southern China (Chan et al., 2011).In addition, the maximum concentration of formaldehyde up to 1.15 mg m -3 have been detected in some renovated shopping malls in Western China (Shang et al., 2016).Tao et al. (2015) found that mass concentrations of formaldehyde ranged from 0.05 mg m -3 to 0.26 mg m -3 as measured in nine underground malls in Xi'an.A recently published article by Tang et al. (2005) also reported that the indoor concentrations of formaldehyde in underground malls generally have exceeded the standard, which may cause malls to be unhealthy.What's more, the characteristics and concentrations of formaldehyde vary widely in diverse indoor environments because of the various types of pollutant emission sources (Tao et al., 2015;Shang et al., 2016).It has been shown that the increasing use of wood composites for decoration, including new wallpaper and wooden floors and ceilings, will inevitably cause deterioration of indoor air quality (Gao et al., 2012;Lv, et al., 2016).So far, numerous indoor air pollution studies in China have focused on the concentration of formaldehyde in indoor public places, but little is known about the influence of other factors on the formaldehyde concentrations in these typical indoor environments.Considering these facts, a comprehensive study about the external factors that affect formaldehyde emission in indoor public places seemed worthy of effort.
Guangxi is one of the most famous Chinese provinces in China for tourism, so the air pollution problems in Guangxi have received significant attention from tourists and the local government.This is important, because many people expend considerable time pursuing indoor activities.When they are recreating or working in the indoors, their exposure to formaldehyde increases dramatically.In the present study, objective measurements were performed in four typical indoor crowded venues that accommodate multiple activities to obtain a much clearer picture of indoor formaldehyde pollution in Guilin, Liuzhou, and Nanning in Guangxi, China.This study focused on (1) the current indoor formaldehyde pollution levels in typical places in Guangxi, China (2) the temperature, humidity, and air change rate effects on indoor formaldehyde concentrations in four typical public places.

Description of the Sampling Locations
Four specific public places including furniture markets, malls, hotels, and restaurants were sampled in the three selected cities of, Guilin, Liuzhou, and Nanning in Guangxi, China.The detailed information and test conditions of the sampling sites, including temperature, humidity and air change rate for each places during the measurement period, are summarized in Table S1.To eliminate the variations between the sampling sites, the same types of places were randomly selected in the relatively prosperous sections of the three cities and these places had not been furnished or refurbished for the previous three years.Furthermore, to ensure the accuracy and reliability of the sample data, the measurements in the furniture markets and malls were conducted from 10:00-14:00 between June 1st and September 30th, 2014.The measurements in hotels and restaurants were conducted from 10:00-14:00 between July 1st and August 30th, 2015.The layout of the measuring points in a typical place in this case study (GM1 as an example) is shown in Fig. S1.The sampling points were evenly distributed and situated more than 0.5 m from the ventilation vents and interior walls.In addition, the sampling was obtained 1.5 m above the floor level to simulate the area in which people would inhale the available air in each location.Similar sampling strategies can be also found elsewhere (Chang et. al., 2017).In each of the public places the measurements lasted from one to seven days depending on the availability of the locale.All the apparatus used in this study had been calibrated before the field tests and the measurements at each sampling points were repeated three times to determine the precision of the measurements.

Experiments and Methods
This study was designed to investigate the formaldehyde concentrations at four typical public places sampled using an active sampler (KC-60, Qingdao Laoshan Electronic Instrument Factory Co. Ltd., China) with bubbling absorption tube.Before sampling, the flow constant of the sampling apparatus was precisely calibrated using a mass flow meter (TSI-4140, TSI Inc., USA), with a sampling range of 0.01 Std L min -1 to 20 Std L min -1 with an accuracy of ± 2%.Once the sampling was completed, the samples were placed in a ziplock plastic bag and then sent to the laboratory for analysis.Formaldehyde samples were determined by phenol reagent spectrophotometry in accordance with the analytical methods prescribed in the Chinese National Standard (GB/T 18204.26-2014).Related analytical method of indoor formaldehyde can be also found in elsewhere (Chang et. al., 2017).Briefly, the formaldehyde was first collected by the use of KC-60 sampler, whose flow rate set at 0.5 L min -1 , with a phenol reagent that was used as the absorbing liquid in bubbling absorption tube for 20 minutes to form azine. Then previously prepared high electrovalent ions were added into the collected samples to affect the reaction of azine for about 15 minutes at the room temperature to form bluishgreen compounds.Finally, the concentrations of formaldehyde in the samples were evaluated using a spectrophotometer (752N, Shanghai optical Instrument Factory Co. Ltd., China) with the characteristic UV peak centered at 630 nm.The values of indoor temperature and indoor relative humidity were continuously recorded using a hygrothermograph (HT-1292, Guangzhou Tailang Instrument Factory Co. Ltd., China).The range of indoor temperature measurement was from 10°C to 60°C with an accuracy of ± 0.5°C, and the range of indoor relative humidity measurement was from 10% to 95% with an accuracy of ± 1%.
The indoor air change rate was estimated using carbon dioxide as a tracer gas and follow-up of its decay as described in ASTM Standard E741 (ASTM, 2006).Briefly, 50 L from a compressed cylinder of carbon dioxide (99%) were initially injected in the middle of a location in order to increase the indoor carbon dioxide concentration.Tracer gas concentrations were typically well mixed after about 30 min in a mechanically ventilated or relatively small (< 1000 m 2 ) space.Fans were used to help with the mixing of the tracer gas in relative larger places (> 1000 m 2 ) that were naturally ventilated.The initial indoor carbon dioxide concentration was measured by multiple ambient air conditions measuring instruments (Testo 435-4, Testo AG, Germany) with the corresponding probe (IAQ 0632 1535) that was uniformly distributed at various locations in the select location.Using this approach, the initial indoor carbon dioxide concentration can be determined precisely to within about 5% of each other after a certain time until the concentration measured by these analyzers increases to approximately 1000 ppm.The changes of carbon dioxide concentrations were continuously monitored in the middle of the room for about 3 h.The range of carbon dioxide measurement was from 0 ppm to 5000 ppm with an accuracy of ± 50 ppm.The air change rate was directly deduced using a semi-logarithmic plot (Dorizas et al., 2015).
where A is the air exchange rate (h -1 ), t is the time (h), and C t and C 0 are the indoor carbon dioxide concentrations at time t and 0, respectively.

Statistical Analyses and Evaluation Standard
In this study, statistical analyses were performed on the measured data using SPSS software (version 23.0), Origin (version 9.1) and Microsoft excel (version 2013).There was no unified limited standard for the formaldehyde concentration.Consequently, WHO guidelines for indoor air quality of selected pollutants (WHO, 2010) was used as a reference to evaluate the formaldehyde concentrations.The maximum limit for formaldehyde concentration was 0.1 mg m -3 (per 30 min).

Quality Assurance and Quality Control
In order to ensure the reliability and accuracy of the results, the quantification of formaldehyde analysis was conducted in strict accordance with the operating procedures described in the standard methods (GB/T 18204.26-2014).Before the analysis of the absorbed samples, an aqueous 1.00 µg mL -1 formaldehyde solution was used as a standard solution for calibration.A formaldehyde concentration calibration curve was established using nine standard formaldehyde concentrations.Meanwhile, the blanks of all the reagents were found to be low in formaldehyde as deducted in the analysis.Strong linear relationships (R 2 > 0.999) were found between the concentrations and responses for formaldehyde.A calibration standard, including routine operational maintenance and standard sample calibration, was run each day with the same analytical procedures as the test samples to ensure the stabilization of the instruments.Meanwhile, parallel samplings were also conducted to ensure the validity of the results.Generally, three parallels analyses of each sample were conducted, which yielded a relative standard deviation less than 5%.

Formaldehyde Concentrations of Indoor Air
The formaldehyde concentrations found in the selected sites during the experimental period are presented in Fig. 1.Based on the measured sampling data, it appeared that the formaldehyde concentrations measured in the furniture markets were generally higher than the other public places except for the furniture market NM3.The over standard rate of formaldehyde concentration (0.1 mg m -3 ) measured in furniture markets was 94%.The highest formaldehyde concentration of 0.313 mg m -3 was detected in the furniture market NF5 (Fig. 1(a)).This was attributed to a large quantity of wood-related products sold in the furniture market.The urea-formaldehyde (UF) adhesives and paint employed in these products contain high levels of formaldehyde, which outgas a large amount of formaldehyde and produce serious indoor formaldehyde pollution (Guo et al., 2014).Mendell et al. (2007) also reported that the paint, latex paint, new furniture and bonded wood can significantly contribute to the content of formaldehyde in the air.In the case of the shopping malls, it was found that 13 out of the 15 sampled sites in the malls produced formaldehyde concentrations that were outside of the acceptable limit (0.1 mg m -3 ) with the over standard rate of 87%.The highest formaldehyde concentration (0.231 mg m -3 ) was measured in LM5 in Liuzhou (Fig. 1(b)).This may have been due to the fact that the malls were crowded with abundant merchandises composed of fabric and leather that might contain formaldehyde-based adhesives.Specifically, it was found that malls with many of wooden shelves had higher formaldehyde concentrations.In addition, the use of various chemical products, that contain a resin leather finishing agent, color fixing agent, and softening agent, used in the leather, bedding, and clothing products were also major sources of formaldehyde that caused a deterioration of the indoor air.Furthermore, there were three hotels: Hotels GH4, LH1, and NH5, where the formaldehyde concentrations exceeded the acceptable limit (0.1 mg m -3 ) with the exceeding standard rate of 20% (Fig. 1(c)).This may have been due to the large quantity of domestic products in the hotels including scattered carpets, plywood panels and different types of floorings that inevitably emit formaldehyde producing high levels of formaldehyde concentration in the air (Liu et al., 2014;Mishra et al., 2015).In addition, the formaldehyde concentrations tested in restaurants were found to conform to the national standard (Fig. 1(d)), which means that the pollution level in restaurants was not as serious as furniture market and malls.Based on the measured data, we considered that the relatively good indoor air quality in restaurants may be partly due to their relatively higher ventilation effectiveness and air exchange efficiency (0.94 h -1 -1.69 h -1 for restaurants) than the other three types of places (0.54 h -1 -1.16 h -1 for furniture markets, 0.66 h -1 -1.26 h -1 for malls, and 0.66 h -1 -1.40 h -1 for hotels).In general, the variation in the formaldehyde pollution levels in the different types of indoor environment may be mainly due to the different types of decorations or merchandises, as well as the various typical pollutants in the four kinds of typical public places.
To intuitively deduce the overall pollution level in the various places, the formaldehyde concentrations in these varied public places in the three selected cities were further categorized and displayed as shown in Table 1.This investigation showed that the mean formaldehyde concentrations tested in furniture markets and malls in Guilin, Liuzhou, and Nanning were all outside of the acceptable limit (0.1 mg m -3 ) with a percentage surplus ranging from 42% to 123% above the accepted concentration limit.The formaldehyde pollution levels in hotels were generally not serious compared to furniture markets and malls that had mean formaldehyde concentrations ranging from 0.075 mg m -3 to 0.082 mg m -3 .Unlike furniture markets, malls, and hotels, the mean formaldehyde concentration found in all the restaurants were generally in the low value range, which indicated that the restaurants in the three selected cities had relatively good indoor air quality.It was further found that the overall mean formaldehyde concentration of the four typical places in descending order was furniture market > mall > hotel > restaurant.In general, the remarkable difference in the formaldehyde pollution levels of the four types of public places may have been due to the contributions of several factors such as different variations of indoor environmental conditions, the various pollutant sources, the position of the emission sources, as well as variations of environment conditions (Parthasarathy et al., 2011;Yu and Kim, 2012).

Analysis of the Distribution of the Formaldehyde in Four Typical Public Places
There is a large quantity of decorations used in the indoor design of buildings and these materials are rich in formaldehyde.This is a major contributing factor that leads to the indoor environmental pollution in many of the sampled venues and complicates amelioration of the quality problem.A selection of boxplots has been included in Fig. 2 that represent the mass concentration of formaldehyde in the various places.These plots provide a graphical representation of the variation in the concentration of formaldehyde in the four typical public places, which showed that there was a complex spatial distribution in the emission characteristics of formaldehyde in the indoor air.Also, the median concentration of formaldehyde, including the minimum and maximum concentrations in the different types of public palaces, are presented in Table S2.This investigation showed that the indoor air formaldehyde concentrations can vary over a wide range within the same type of locale, particularly in the furniture markets in Nanning (0.015 mg m -3 -0.473 mg m -3 ), the malls in Liuzhou (0.058 mg m -3 -0.337 mg m -3 ), the hotels in Guilin (0.009 mg m -3 -0.195 mg m -3 ), and the restaurants in Nanning (0.001 mg m -3 -0.127 mg m -3 ).Moreover, it was found that all the median formaldehyde concentrations in the furniture markets (except for the furniture market NF3) nearly exceeded the acceptable limit (0.1 mg m -3 ).Likewise, it also can be seen that the top onefourth (quartile) of the formaldehyde concentrations in the furniture markets and malls in three selected cities (except for the mall NM1) nearly exceeded the standard limit (0.1 mg m -3 ), which further indicated that formaldehyde pollution level in these locations was a serious concern.Considering the data in this study, it was suspected that the variations in the formaldehyde concentrations in the different types of indoor environments can be partly attributed to the widespread pollution sources, including the quantity of decorations and the types of materials that have been identified as the important sources of pollution in the indoor environments.The wide distribution of these materials throughout the location complicated the contamination scenario (Hwang et al., 2011;Li et al., 2013).

Formaldehyde Pollution Levels in Comparison with Other Studies
For comparison, the pollution levels of formaldehyde in different public places in various cities as reported in the literature (Ding et al., 2002;Wang, 2004;Zhang et al., 2005;Jiang et al., 2010;Cui et al., 2015;Li and Yang, 2015;Zhong et al., 2015) after the year 2002 are summarized and listed in Table 2.It can be seen from these data that the overall mean formaldehyde concentrations in furniture markets determined in this reported research were much higher than those reported in Weifang.In malls, the overall mean formaldehyde concentrations measured in Guangxi were all higher than those measured in Xi'an and Tianjin.This may have been due to the unique climate of Guangxi, which has relatively higher ambient temperature and humidity conditions in the summer (Lu et al., 2016) that may produce higher formaldehyde concentrations in the air in the selected test venues.In addition, the air samples collected during 10:00-14:00 between June 1st and September 30th may have had higher formaldehyde content, because this time period represents the highest annual temperature for the year in this locale.Furthermore, formaldehyde measurements in the hotels in Guangxi, Tianjin, and Suzhou were also close to the allowable limits (0.1 mg m -3 ).The indoor formaldehyde concentrations observed in this study suggests that the indoor formaldehyde concentration levels in the four types of public places has always been a serious problem in China.Combining documented literatures data (Parthasarathy et al., 2011;Gao et al., 2012;Yu et al., 2012;Liu et al., 2014;Mishra et al., 2015) with some of the conclusions obtained in this study, we conclude that the widespread existence of indoor formaldehyde pollutant sources including decorations, merchandise, wood composites, chemical auxiliaries, etc., were the primary causes for the deterioration of the indoor air quality.However, we cannot ignore the other indoor environmental factors in this situation including temperature, humidity, and air change rate that are also important factors affecting the variation of formaldehyde concentrations in the different types of indoor environments.

The Environmental Factors Affecting on Formaldehyde Concentrations
To further determine the influence of the various factors on the emission behaviors of formaldehyde in four kinds of typical places, three indoor environmental factors, including temperature, humidity, and air change rate influencing on  indoor formaldehyde were systematic analyzed by combining the 4 sets of experimental data obtained in this survey.The correlation matrix of the formaldehyde concentrations in the different types of public places and the relative influential factors are shown in Table 3. Applying the Pearson test, for each type of place we found statistically significant differences (p < 0.01) between the indoor formaldehyde concentrations and temperature, humidity, and air change rate.In addition, strong associations (r > 0.80) were found between the formaldehyde concentration and these factors in the indoor environments that further demonstrated that temperature, humidity and air change rate had a significant impact on the indoor formaldehyde pollution level.

Temperature Effects on Indoor Formaldehyde Concentrations
The temperature of the various public locations was found to directly affect the concentration of formaldehyde in the indoor air as shown in Fig. 3, which demonstrates that the formaldehyde concentrations emitted from four typical places had similar emission characteristics.Strongly positive correlations (r > 0.82, p < 0.01) were found between  the formaldehyde concentration and indoor temperature in the four public places, which further indicated that the formaldehyde concentrations were significantly influenced by the temperature of the room and low temperatures helped to decrease indoor formaldehyde concentrations.The rise of formaldehyde emissions in the typical places appeared to result from elevated temperatures that accelerated the adsorption or desorption of low weight formaldehyde molecules on articles in the room, which increased the release of formaldehyde in to the air.
Some studies have illustrated that the formaldehyde concentrations in indoor places were higher in the summer, with relative higher temperatures, but lower during the winter (Guo et al., 2014;Lu et al., 2016), which provides some corresponding evidence that the formaldehyde concentrations in different public places were significantly influenced by indoor temperature.It is worth noting that Guangxi has a unique climate with relatively higher temperature climatic conditions in summer (Lu et al., 2016).As a result, the use of central air conditioning (CAC) in indoor environment to reduce indoor temperature seemed to be an effective way of alleviating the indoor formaldehyde pollutants.

Humidity Effects on Indoor Formaldehyde Concentrations
The effect of the relative humidity and absolutely humidity on indoor formaldehyde concentrations in the different public places is presented in Figs. 4 and 5, respectively.Based on these results it was found that the indoor formaldehyde concentrations in the four venues were more strongly correlated (r > 0.85, p < 0.01) with absolutely humidity than with relative humidity (r < 0.83, p < 0.01), indicating that absolutely humidity had more of an influence on the indoor formaldehyde concentrations.Similar conclusions obtained in other studies (Liang et al., 2015;Liang et al., 2016) provide further support for the findings of present study.A hypothesis similar to that previously reported may explain how high levels of urea-formaldehyde adhesives used in the pollution sources could depolymerize through hydrolysis in the presence of free water so that the amount of available free water would determine the amount of hydrolysis.However, the relative humidity can vary with temperature even when the molar amount of free water remains the same in the air.The absolutely humidity, by definition, represents absolute quantity of free water in the air, which serves as a better determination of the amount of available free water.Consequently, absolutely humidity may be a better measure of the effect of humidity on indoor formaldehyde emissions and also explains why there is a strong positive correlation between the formaldehyde concentration and absolutely humidity in the four types of public places (Liang et al., 2015).
As is well known, Guangxi has unique climate conditions with high humidity averaging at 70-80 percent throughout the year (Lu et al., 2016).Considering the literature reports about formaldehyde concentration in indoor air and humidity from similar studies (Harving et al., 2010, Liang et al., 2016), it can be concluded that control of indoor humidity is an effective way to control indoor formaldehyde concentration.Therefore, use of a desiccant cooling system (DECS) indoors in humid weather conditions appears to be an effective and practical method for removing indoor air pollution of formaldehyde.

Air Change Rate Effects on Indoor Formaldehyde Concentrations
The air change rate was found to effect indoor formaldehyde concentrations in the four types of typical places as shown in Fig. 6, which showed that indoor formaldehyde concentrations gradually decreased as the air change rate was increased.Strongly negative correlations (r > 0.80, p < 0.01) were found between the indoor formaldehyde concentration and air change rate, which demonstrated that the formaldehyde concentrations were significantly influenced by the indoor air change rate.These results agree with similar correlations obtained in other studies (Hult et al., 2015;Ho et al., 2016;Meng and Hong, 2017), which further proves that improving the ventilation in indoor environments is an effective way to control formaldehyde indoor air concentrations.
The air change rate depends primarily on the design of the building and the ventilation behavior.Typical indoor locations generally have poor air quality because of their enclosed structure and poor natural ventilation.A closed and insufficiently ventilated indoor environment will have higher pollutant levels due to a lack of fresh air.Also, a single or irrational mode of ventilation would directly result in a lower air change rate and a high level of indoor formaldehyde concentration.Hence, an appropriate ventilation strategy, which increases the mechanical ventilation

CONCLUSIONS
This work was initiated to determine the formaldehydepollution level in furniture markets, malls, hotels, and restaurants in three cities, which represent the typical public indoor environment in Guangxi.Formaldehyde concentrations measured in the furniture markets and malls were generally higher than the accepted standard (0.1 mg m -3 ) by 94% and 87%.The four types of venues, ranked by overall mean formaldehyde in descending order, were: furniture market (0.193 mg m -3 ) > mall (0.147 mg m -3 ) > hotel (0.079 mg m -3 ) > restaurant (0.058 mg m -3 ).The formaldehyde air pollution at these locations in Guangxi remains a serious concern because high levels of formaldehyde can induce health problems in staff and guests.We concluded that the widespread indoor air pollution at each location resulted from the use of decorations, which inevitably emit formaldehyde.Therefore, the most effective way of controlling formaldehyde levels in these environments is to reduce merchandise, household products, and decorations that contain high levels of ureaformaldehyde adhesives.In addition, storing formaldehydecontaining products in a well-ventilated area can produce this desired result.Furthermore, strong associations (r > 0.80, p < 0.01) were found between aerial formaldehyde concentrations and indoor environmental factors, which  demonstrated that indoor temperature, humidity, and air change rate have a significant impact on the formaldehydepollution level indoors.Based on this, it appears that artificially controlling indoor environmental conditions (temperature, humidity, and air change rate) is the most effective method for limiting the formaldehyde levels.In particular, an appropriate ventilation strategy that employs the reasonable combination of mechanical and artificial ventilation in an indoor environment can be one of the most effective and efficient ways to reduce the indoor formaldehyde-pollution levels.Further research that might establish a mathematical model to predict indoor formaldehyde concentrations resulting from the presence of various building materials is needed.

Fig. 1 .
Fig. 1.Formaldehyde concentrations in (a) furniture markets, (b) malls, (c) hotels, and (d) restaurants in three select cities.The error bars represents the standard deviation.

Fig. 2 .
Fig. 2. Mass concentration of indoor formaldehyde in (a) furniture markets, (b) malls, (c) hotels, and (d) restaurants in three selected cities in the present study, box-whisker plots represent the 5th, 25th, 50th, 75th and 95th percentiles, as well as the mean.
is significant at the 0.01 level (two-tailed).

Table 1 .
Formaldehyde concentrations in locales in four types of typical public places.

Table 2 .
Formaldehyde concentration in different type of public places reported in earlier Chinese studies.Bold numbers mean the values exceeded the limits (0.1 mg m -3 ).

Table 3 .
Correlation matrix among formaldehyde concentrations in different typical places and affecting factors.