E ﬀ ects of Indoor Plants on the Physical Environment with Respect to Distance and Green Coverage Ratio

: Few studies have conducted experiments in daily living environments to examine the e ﬀ ects of indoor plants on objective aspects of the physical environment. This study examined the e ﬀ ects of plant distance and green coverage ratio on the objective physical environment and subjective psychological perceptions, along with the correlation between the objective physical environment and subjective psychological perceptions regarding indoor plants. A randomized control trial of plant distance and green coverage ratio was conducted in a room located in the basement of a university building in Taiwan. Aspects of the objective physical environment were measured using air quality detectors. Subjective psychological perceptions were evaluated based on the questionnaire responses of 60 undergraduates. The results revealed that (1) regardless of number of plants, the closer the plant, the higher the CO 2 level; (2) more indoor plants resulted in higher CO 2 and humidity and lower PM 2.5 , PM 10 , and temperature; and (3) the lower the levels of ﬁne and suspended particles in the air were, the stronger were the feelings of preference, naturalness of the environment, and pleasure in participants. Indoor plants that can regulate indoor air quality and microclimates without consuming energy warrant greater attention and wider application.


Introduction
People in contemporary society spend 80% to 90% of their time indoors every day [1]. Thus, having a comfortable indoor environment with favorable indoor air quality is imperative, as indicated by the increasing prevalence of sick building syndrome [2]. In the United States alone, roughly 27 million office workers are at risk of sick building syndrome, and 30% of new buildings worldwide are associated with indoor air pollution problems [3]. Moreover, indoor air pollution is generally 2 to 5 times worse (sometimes up to 100 times worse) than outdoor air pollution [4]. This may be one of the reasons that people in contemporary society are experiencing increasingly severe physical and mental health problems, along with declining general wellbeing [5].
By contrast, exposure to nature is considered beneficial to physical and mental health [6]. This is why people often grow plants indoors to improve the quality of their living environment and workspace [7], because plants represent nature [8,9]. Empirical studies have also demonstrated that indoor plants have positive effects on physical and mental health as well as general wellbeing [10,11]. Therefore, the beneficial effect of indoor plants on public health is a topic requiring exploration.
Research has shown that having even a few indoor plants is beneficial to the general wellbeing and physical and mental health of humans. For example, covering 6% of an indoor floor area with indoor plants can elicit significantly stronger feelings of preference, comfort, and friendliness in students and reduce their misbehavior and number of sick leave days [12]. A study conducted in real offices showed that the presence of plants has a significant effect on reducing short-term sickness absence, compared to Hypothesis 4 (H4). Air quality, temperature, and humidity correlate to novelty, perceived naturalness, preference, emotions, and environmental comfort.

Research Design
This study, which was part of a series of experiments for a research project, conducted a randomized controlled trial. To examine the effects of indoor plants on the physical environment, we constructed control conditions and experimental treatments. The order of the experimental treatments was randomly decided. To examine the effects of indoor plants on the perceptions of participants, participants were randomly assigned to different experimental treatments. All participants gave informed consent for inclusion before they underwent the experiment. The study was conducted in accordance with the Declaration of Helsinki of 1975, and was approved by the Ministry of Science and Technology in Taiwan, which did not require this study to be submitted to the Ethics Committee.

Experimental Setup
The experiment for this study was conducted in a room located in the basement of a technology university in central Taiwan. The room (length 4.0 m × width 3.0 m × height 3.85 m) was newly constructed in October 2018. During the experiment, the windows in the room were kept shut, the air-conditioner was switched off but the air conditioning vents were not blocked, and the curtains were drawn to avoid distractions from the view of the landscape outside. The experimental environment tended to be more like an actual living or working space rather than a completely isolated laboratory unit. Except for participants, no one entered or exited the room during the experiment. However, when the participants opened the door and entered and exited the room, the physical environment of the room changed.
The Rural Development Administration of South Korea recommends installing one large potted plant and one small potted plant per 6 m 2 floor area in a room [29]. US studies have also recommended placing at least one 6-in (15.24 cm) potted plant per 9 m 2 floor area in a room [30] to purify indoor air. Since the room in which our experiment was conducted had a floor area of 12 m 2 at least two large potted plants (diameter of 24 or 30 cm) and two small potted plants (diameter of 18 cm) should be installed to improve the indoor air quality. For simplicity, this study selected only large potted plants (pot diameter 37 cm and plant height of around 1.3 m) because the eye of a sitting person is at a height of approximately 1.1 m from the ground [31]. The indoor plant used was the evergreen Radermachera hainanensis Merr. a shade-bearing tree whose brightly colored leaves give the plant its ornamental value [32]. Though this tree has become popular for indoor use in Taiwan in recent years, its ability to improve air quality and its influence on people's responses has never been examined. Plants tested in experiments for their effects should be used as plants are normally used in daily living. The plant pots were made of plastic. The growing medium was potting soil. Since we viewed the plants, pots, and potting soil as a whole unit, we did not distinguish the possibly different influences of the plants, pots, and potting soil.
The experimental manipulation to achieve a low green coverage ratio (3.00 %) involved placing one large potted plant against the wall farthest from the door at the midpoint between two corners of the room. The experimental manipulation to achieve a high green coverage ratio (8.83 %) involved placing three large potted plants against the wall so that the tree canopies of all three plants were touching. The low and high green coverage ratios were intentionally limited because previous research has found that even a few indoor plants are beneficial to human health [12,14]. Further, we followed the recommendations for the minimum number of indoor plants for improving air quality [29,30]. Table 1 details the plant distance and green coverage ratio of the experimental treatments. Before each of the experimental treatments, a control condition was conducted with no plants in the room. The participants of this experiment were 60 students studying in daytime programs at a technology university in central Taiwan. For all experiments, participants sat on a chair placed in the center of the room at a distance of 1.5 m from the potted plant(s) (Figure 1). In the experiment, participants were asked to sit down, observe the surrounding environment for 1 min, and ignore the two measuring instruments as much as possible. After observing the room, participants began to fill out the questionnaire, and were given as much time as they required, provided that they spent at least 5 min on it. Further, we followed the recommendations for the minimum number of indoor plants for improving air quality [29,30]. Table 1 details the plant distance and green coverage ratio of the experimental treatments. Before each of the experimental treatments, a control condition was conducted with no plants in the room. The participants of this experiment were 60 students studying in daytime programs at a technology university in central Taiwan. For all experiments, participants sat on a chair placed in the center of the room at a distance of 1.5 m from the potted plant(s) (Figure 1). In the experiment, participants were asked to sit down, observe the surrounding environment for 1 minute, and ignore the two measuring instruments as much as possible. After observing the room, participants began to fill out the questionnaire, and were given as much time as they required, provided that they spent at least 5 minutes on it.

Objective Measurement of the Physical Environment
The physical environment was objectively measured using two indoor air quality detectors, which were both placed slightly to the left of the chair. The two detectors were placed 1.1 m above the ground and 0.25 and 1.5 m away from the bottom of the potted plant(s), respectively. The indoor air quality detectors used in this study (iAeris 14, Kaoten Scientific Co., Ltd., Kaohsiung, Taiwan) can measure changes across eight aspects of the physical environment: temperature, humidity, carbon monoxide (CO), carbon dioxide (CO 2 ), coarse particulate matter (PM 10 ), fine particulate matter (PM 2.5 ), formaldehyde (HCHO), and total volatile organic compound (TVOCs) levels ( Table 2 for specifications). The air quality detectors had quality control and quality assurance, which were calibrated again before the experiment. The air quality detectors were switched on throughout the experiment and uploaded the data measurements to the cloud every 6 min. Therefore, there was no repetition in the measurements of the physical environment.

Subjective Measurement of Psychological Perceptions
This study employed a questionnaire to measure the subjective perceptions of the participants. The questionnaire contained six sections: (1) preference, (2) environmental comfort, (3) perceived naturalness, (4) novelty, (5) the pleasure-arousal-dominance (PAD) emotional state scale, and (6) participant demographic information. The participants selected their answer on a 9-point Likert scale on which a higher score indicates a greater level of agreement with the questionnaire item. Participant demographics consisted of sex, year of birth, college, department, and year of study.

Preference
Preference reflects an immediate and direct response to or assessment of the environment in general [33,34] and can be either task-dependent, which is related to a specific activity, or task-independent, which is unrelated to any particular activity [35]. The survey of preference for indoor plants in this study comprised two items: "mere liking," which is task-independent, and "whether (the participants) liked to study in the space," which is task-dependent [36,37]. The total average score for these two items constituted the index score for the preference.

Environmental Comfort
Few studies have investigated the effects of indoor plants on comfort level. Nevertheless, one literature review derived four dimensions of indoor environmental quality: thermal comfort, visual comfort, acoustic comfort, and indoor air quality [38]. Therefore, the environmental comfort examined in this study also covered these four items: How do you evaluate the indoor thermal environment? How do you evaluate the indoor visual comfort? How do you evaluate the indoor acoustic comfort? How do you evaluate the indoor air quality [39]? The index score for the Sustainability 2019, 11, 3679 6 of 19 environmental comfort was obtained by calculating the average of the total scores for these four items. A higher score indicates greater comfort.

Perceived Naturalness
Perceived naturalness refers to the subjective perception of the closeness of a physical environment to a natural state [40,41]. Higher levels of perceived naturalness of a landscape or environment indicate that a landscape or environment is perceived as closer to its natural state [42]. Ho and Chang [43] developed a 16-item perceived naturalness scale consisting of two constructs: richness of natural elements and environmental wildness. The items used in this study for perceived naturalness were also adapted from other studies to comprise a total of three dimensions (naturalness, richness of natural elements, and environmental wildness), each of which consisted of two items (e.g., Do you think this environment is natural? Is this environment full of vitality? Is this environment without a trace of human destruction?). A higher score indicates a greater level of naturalness. The index score for the perceived naturalness was calculated by first determining the composite scores of the three dimensions (i.e., the total score of all items in each dimension divided by the number of items) and then averaging the composite scores.

Novelty
A novelty scale suitable for this study has not yet been developed. Therefore, this study adopted two related scales that have been used in Taiwan and overseas. The first scale was obtained from Lee and Crompton [44], who argued that novelty has four dimensions: change from routine, thrill, surprise, and boredom alleviation. The second is the familiarity scale developed by Cheng, Shen, and Chon [45], given that familiarity is the opposite of novelty [46]. Cheng et al. [45]. classified familiarity of a place into four categories based on Baloglu [47]: overall familiarity, having seen it, having the ability to recognize it, and having visited it. The novelty items in this study were adapted from the two aforementioned scales and adjusted to include two items from each of the four dimensions specified by Lee and Crompton [44] and the three types of familiarity (not including overall familiarity) proposed in Cheng et al. [45], which were reverse-worded. Therefore, the novelty items proposed in this study comprised a total of five dimensions and eleven questions (e.g., There are new things in this environment that I can explore. This environment feels a bit dangerous to me. This environment makes me feel a little surprised. This environment can reduce my boredom. I have never seen this environment or a similar environment.). A higher score denotes greater novelty. The index score of the novelty was calculated by first determining the composite scores of the five dimensions (i.e., the total score of all items in each dimension divided by the number of items) and then averaging the composite scores.

Emotions
Emotion is considered to be an innate human quality, as evidenced by the emotional expressions of newborn infants [48,49]. Emotions are widely believed to be the fundamental factor influencing environmental perception [50][51][52]. This study adopted the PAD emotional state scale [53], which consists of three dimensions: pleasure-displeasure, arousal-nonarousal, and dominance-submissiveness. Each dimension has six adjectives (e.g., "happy-unhappy; excited-calm; in control-cared for"), and the scores on the scale ranges from 4 to −4. The scores of the PAD scale were calculated by determining the composite scores of the three dimensions, which represent the average score for each adjective. A higher score denotes a higher level of perception toward the dimension.

Statistical Analyses
The statistical analyses were conducted using IBM SPSS Statistics 22.0 (IBM, Armonk, NY, USA). The data on the physical environment and the psychological perception were first analyzed for their normality and homogeneity of variance, which are the premise of parametric statistical tests such as analysis of variance [54]. The data on the physical environment failed to pass the tests of normality and homogeneity of variance and were, thus, analyzed using nonparametric statistical tests for detecting difference between conditions. The data on the psychological perceptions passed the tests of normality and homogeneity of variance and were analyzed using parametric statistical tests for detecting difference between conditions. In addition, when the focus of interest was the difference within the same condition (controls or treatments), tests with related samples were conducted (e.g., the Wilcoxon tests). When the focus of interest was the difference between different conditions, tests with independent samples were conducted (e.g., the Mann-Whitney U test) [54]. Since the indicators of the difference in the physical environment and the psychological perceptions were the means of the measurement data, the means were below the detection resolution of the air quality detector and the digits of the scale, respectively.

Control Conditions
The control conditions were that the room did not have any plants inside. To match the total number of the data points of the control conditions to that of the experimental treatments, 30 data points from the physical environments were randomly selected for the control condition before each of the experimental treatments. The results showed that although the room had been recently constructed, the air pollution levels of the control conditions were still much lower than the standards mandated under the Indoor Air Quality Act (Table 3). Newly constructed rooms and/or buildings usually have high air pollution levels because of the Volatile Organic Compounds (VOCs) emitted from the building materials such as paint, flooring, caulking, sealants, and engineered wood [55]. Furthermore, there were some significant differences in the physical environment between 0.25 m and 1.5 m, as indicated by the results of the Wilcoxon tests with two dependent samples using the physical environmental factors as dependent variables and the distance as the independent variable. For the control condition of the high green coverage ratio treatment, PM 2.5 and PM 10 at 1.5 m distance were significantly greater than at 0.25 m, while CO 2 , HCHO, TVOC, temperature, and humidity at the 0.25 m distance were significantly greater than at the 1.5 m distance (Table 4). However, for the control condition of the low green coverage ratio treatment, only the temperature at 0.25 m was significantly greater than at 1.5 m (Table 5).

Difference between High Green Coverage Ratio Treatment and Its Control
The results of the Mann-Whitney U tests with two independent samples using the physical environmental factors as dependent variables and the condition as the independent variable showed that PM 2.5 , PM 10 , TVOC, and temperature were significantly greater for the control than for the high green coverage ratio treatment at 0.25 m, while CO 2 , HCHO, and humidity were significantly greater for the high green coverage ratio treatment than for the control at 0.25 m (Table 6). Further, the results of the Mann-Whitney U tests with two independent samples using the physical environmental factors as dependent variables and the condition as the independent variable showed that TVOC and temperature were significantly greater for the control than for the high green coverage ratio treatment at 1.5 m, while CO 2 and humidity were significantly greater for the high green coverage ratio treatment than for the control at 1.5 m (Table 7).

Difference between Low Green Coverage Ratio Treatment and Its Control
The results of the Mann-Whitney U tests with two independent samples using the physical environmental factors as dependent variables and the condition as the independent variable showed that PM 2.5 and PM 10 were significantly greater for the control than for the low green coverage ratio treatment at 0.25 m, while CO 2 , HCHO, temperature, and humidity were significantly greater for the low green coverage ratio treatment than for the control at 0.25 m (Table 8). Further, the results of the Mann-Whitney U tests with two independent samples using the physical environmental factors as dependent variables and the condition as the independent variable showed that PM 2.5 and PM 10 were significantly greater for the control than for the low green coverage ratio treatment at 1.5 m, while CO 2 , temperature, and humidity were significantly greater for the low green coverage ration treatment than for the control at 1.5 m (Table 9).

Difference between High and Low Green Coverage Ratio Treatments
The results of the Mann-Whitney U tests with two independent samples using the physical environmental factors as dependent variables and the experimental treatment as the independent variable showed that humidity was significantly greater for the low green coverage ratio treatment than for the high green coverage ratio treatment at 0.25 m (Table 10). Further, the results of the Mann-Whitney U tests with two independent samples using the physical environmental factors as dependent variables and the experimental treatment as the independent variable showed that CO 2 and humidity were significantly greater for the low green coverage ratio treatment than for the high green coverage ratio treatment at 1.5 m (Table 11).

Difference between Different Distances
The results of the Wilcoxon tests with two dependent samples using the physical environmental factors as dependent variables and the distance as the independent variable showed that HCHO and temperature were significantly greater at 0.25 m than at 1.5 m for the high green coverage ratio treatment, while PM 2.5 and PM 10 were significantly greater at 1.5 m than at 0.25 m for the high green coverage ratio treatment (Table 12). Further, the results of the Wilcoxon tests with two independent samples using the physical environmental factors as dependent variables and the distance as the independent variable showed that CO 2 , HCHO, and temperature were significantly greater at 0.25 m than at 1.5m for the low green coverage ratio treatment, while PM 2.5 and PM 10 were significantly greater at 1.5 m than at 0.25 m for the low green coverage ratio treatment (Table 13).

Participants' Demographic Information
This study had a total of 60 participants (12 men; 48 women), who had an average age of 20.92 years (SD = 1.44). A total of 30 participants (8 men; 22 women) were randomly assigned to receive the high green coverage ratio treatment, and 30 participants (4 men; 26 women) were randomly assigned to receive the low green coverage ratio treatment. Next, a chi-square test was performed with sex, college, department, and year of study as the dependent variables and experimental treatment as the independent variable. The results revealed no significant differences in sex, college, department, and year of study (p > 0.05). Additionally, the results of the independent sample t test revealed a nonsignificant difference between the experimental groups (p > 0.05) in terms of age.

Scale Reliability
This study adopted questionnaires for measuring preference, environmental comfort, perceived naturalness, novelty, and PAD emotional state to investigate the effects of different experimental treatments on the subjective psychological perceptions of participants to examine hypothesis 3. Table 14 provides the reliability of each scale. The Cronbach's α of the environmental comfort scale was 0.67, which is slightly less than ideal. The Cronbach's α of all other scales was greater than 0.728, indicating favorable internal consistency reliability [56].

Difference in the Participant Perceptions between Different Experimental Treatments
The results of the independent sample t test revealed a nonsignificant difference regarding preference (index score) and environmental comfort (index score) between the experimental treatments (p > 0.05), and the difference contained a zero in the 95% confidence interval for the difference (Table 15). One-way multivariate analyses of variance (MANOVA) using the perceived naturalness, novelty, PAD emotions as dependent variables and the experimental treatment as the independent variable revealed that the four variables of perceived naturalness (3 composite scores and index score), six variables of novelty (5 composite scores and index score), and three variables of the PAD scale (3 composite scores) were not significant (p > 0.05), indicating no significant difference between the experimental treatments (Table 16).
Because more than one dependent variable was examined, MANOVA was used since it can simultaneously consider the relationship of several dependent variables, particularly between the composite scores and the index scores in this case, to test whether a significant difference exists in the experimental treatment and analyze whether the experimental treatment exhibits significant differences across the individual dependent variables. However, analysis of variance (ANOVA) hypothesizes that no relationship exists between the dependent variables, and analyzes each individual dependent variable. By contrast, MANOVA is a statistical test of all dependent variables based on their optimal linear combination [57].

Correlations between Objective Factors of the Physical Environment and Subjective Psychological Perceptions
Because the participants sat at a distance of 1.5 m from the plants, their subjective psychological perceptions were subject to a correlation analysis with the objective factors of the physical environment measured at a distance of 1.5 m from the plants. Given that the focus of interest was the relationships between the participants' subjective perceptions and the physical environment with respect to the plants, the correlation analysis used the data on both the low and high green coverage ratio treatments. The results are as follows. (1) Preference exhibited a significantly weak negative correlation with PM 2.5 and PM 10 . (2) Environmental comfort exhibited no significant correlation with air quality, temperature, and humidity. (3) Perceived naturalness exhibited a significantly weak negative correlation with PM 2.5 and PM 10 . (4) The novelty index score exhibited a significantly weak negative correlation with TVOC level. (5) Change in routine in novelty exhibited a significantly weak negative correlation with TVOC level. (6) Pleasure in the PAD scale exhibited a significantly moderate negative correlation with PM 2.5 and a significantly weak negative correlation with PM 10 (Table 17). Note: **(p < 0.01) indicates significance (two-tailed); *(p < 0.05) indicates significance (two-tailed); c denotes that the factor was not calculable because at least one of the variables was a constant. N denotes sample size.

Discussion
The first hypothesis (H1) of this study postulates that different plant distances correspond to different effects on the objective physical environment in terms of air quality, temperature, and humidity level. The results showed that H1 could not be rejected. When the high green coverage ratio treatment was compared to its control, the significantly greater CO 2 , TVOC, and humidity at 0.25 m than 1.5 m were no longer significant. This may because the participants siting at 1.5 m exhaled CO 2 and TVOC that was purified by the 3 plants via their respiration. However, the change in humidity was difficult to explain. It was expected that the humidity would be greater when closer to the plants than when farther away due to the evapotranspiration of the plants. Nevertheless, some plants can reduce the humidity in the air [58]. When the low green coverage ratio treatment was compared to its control, PM 2.5 and PM 10 were significantly greater at 1.5 m than at 0.25 m, while CO 2 and HCHO were significantly greater at 0.25 m than at 1.5 m. This may because (1) PM 2.5 and PM 10 were reduced by the single plant via its respiration [59], (2) the plant itself exhaled CO 2 , though photosynthesis also reduced CO 2 , and (3) HCHO was released by the newly painted wall. Although the control conditions of the two treatments, which were different from each other, made the examinations of the effects of the plants at different distances difficult, it was clear that regardless of number of plants, the closer the plant, the higher CO 2 level. It is also suggested that closer to indoor plants there was less particulate matter.
The second hypothesis (H2) of this study postulates that the effects on the objective physical environment in terms of air quality, temperature, and humidity level would differ when the distance from the plant remains the same but the green coverage ratio varies. The results partially supported H2. Regardless of the distance from the plant (1.5 m and 0.25 m), CO 2 and humidity were significantly greater for the high green coverage ratio treatment than for its control, while TVOC and temperature were significantly greater for the control than for the high green coverage ratio treatment. Further, regardless of the distance from the plant, CO 2 , temperature, and humidity were significantly greater for the low green coverage ratio treatment than for the control, while PM 2.5 and PM 10 were significantly greater for the control than for the low green coverage ratio treatment. Therefore, regardless of the distance from the plant, CO 2 and humidity were significantly greater for the presence of indoor plants (3 and 1 plants) than for the absence of indoor plants. This may because the plant exhaled CO 2 and transpirated water. In addition, it suggested that three indoor plants could reduce temperature at both 1.5 m and 0.25 m because of transpiration, while one indoor plant could not reduce the temperature at both 1.5 m and 0.25 m. However, humidity was greater for the low green coverage ratio treatment than for the high green coverage ratio treatment at both 1.5 m and 0.25 m, which was difficult to explain. It was expected that more plants produce greater humidity unless the plants can reduce humidity [58].
The third hypothesis (H3) of this study postulates that the effects on subjective psychological perceptions in terms of novelty, preference, perceived naturalness, emotions, and environmental comfort would differ when the distance from the plant remains the same but the green coverage ratio varies. The results did not support H3. Therefore, the green coverage ratio (8.83% or 3.00%) did not influence the subjective psychological perceptions of participants. This result differs from Han [12], who reported that a 6% green coverage ratio could elicit significantly stronger feelings of preference and comfort in students. However, this study adopted an experimental method to investigate the preferences (task-dependent and task-independent) and environmental comfort (four items) of university students following brief exposure to R. hainanensis Merr., whereas Han [12] conducted a field quasi-experiment to examine the preference (one item) and comfort (one item) of students from junior high schools after prolonged exposure to Cinnamomum kotoense Kanehira et Sasaki. Therefore, a direct comparison of our study with that of Han [12] may not appropriate. In addition, this study discovered that the participants' feelings of novelty were unaffected, irrespective of the number of potted plants. Hence, the finding reported in other studies that only a few indoor plants benefit peoples' physical and mental health as well as overall wellbeing [12,14,16,60] is unlikely attributable to the effect of novelty. In future investigations of the effects of indoor plants on participants' subjective psychological perceptions, researchers may need to further increase the green coverage ratio to identify possible effects.
The fourth hypothesis (H4) of this study postulates that air quality, temperature, and humidity correlate with novelty, perceived naturalness, preference, emotions, and environmental comfort. The results partially supported H4. Preference, perceived naturalness, and pleasure exhibited a significant negative correlation with PM 2.5 and PM 10 levels. Therefore, the lower levels of fine and suspended particles in the air, the stronger the feelings of preference, naturalness of the environment, and pleasure in participants. Furthermore, both change in routine in the novelty and the index score for novelty exhibited a significant negative correlation with TVOC level, suggesting that the lower levels of TVOC in the air, the stronger the feelings of novelty and change from routine in participants. This may because participants spend much time in TVOC-polluted rooms [61] and likely have become accustomed to such environments. Finally, environmental comfort and objective physical environmental factors were not significantly correlated, possibly because the air quality, temperature, and humidity level of the experimental room did not vary considerably and were within standard ranges defined by the Indoor Air Quality Act (Table 3).
This study examined the effects of indoor plants on both the physical environment and the psychological perceptions, manipulating their distances and green coverage ratios. This made the research design relatively complicated. A limitation of this study was that the participants were situated only at 1.5 m but not at 0.25 m from the plant, which resulted in an unbalanced design. The air quality detector at 1.5 m would be more affected by its proximity to the participant than the plant. Although any person in the room would emit heat, VOCs, CO 2 , and would likely re-suspend particulate matter, this was not the case in this study, except for CO 2 . Future studies may adopt more straight forward approaches than this one. Moreover, the analyses of the data on the physical environment relying on the nonparametric statistical tests could not take into account the effects of both the plant distances and amount simultaneously. Therefore, the effects of the plant distances and green coverage ratios on the physical environment were analyzed separately. When combining the pieces of information together, the results showed a clearer picture: (1) regardless of number of plants, the closer the plant, the higher the CO 2 level; (2) the major effects of more indoor plants were greater CO 2 and humidity; and (3) the secondary effects of more indoor plants were less PM 2.5 and PM 10 as well as lower temperature. Furthermore, given that the experiments of this study were conducted in an actual environment, some variables such as weather could not be fully controlled. It rained for two days and one day during the high and the low green coverage ratio treatment, respectively, which may influence the physical environment (e.g., humidity) and the psychological perceptions (e.g., emotions) [62]. More replication studies are needed.
More studies on the physiology of R. hainanensis Merr are necessary. How much water the plant needs for photosynthesis, how much it transpirates, how much CO 2 its exhales during its breathing, and how much it absorbs during its photosynthesis remain unknown. Laboratory studies and field studies are both needed.

Conclusions
This study was the first to verify that the distance from indoor plants and their green coverage ratio in a real setting can significantly influence the physical environment. Simply placing one or three large pots of R. hainanensis Merr. in a room with a floor area of 12 m 2 can increase the humidity levels and reduce PM 2.5 and PM 10 . Further, three large pots of R. hainanensis Merr. can reduce temperature more significantly than one large pot. These findings bridge the gap in research regarding experiments that mostly involve highly concentrated pollutants in small and enclosed fumigation chambers, which differ from people's daily living environments. Furthermore, the literature still lacks investigations of the distance and green coverage ratio of indoor plants. This study also demonstrated findings that only a few indoor plants benefit people's physical and mental health and wellbeing are probably not attributable to the effect of novelty, and further revealed that even a few pots of indoor plants in a minimally polluted room, which meets the Indoor Air Quality Act, can still effectively improve the air quality and environmental comfort. Therefore, indoor plants that can regulate indoor air quality and microclimates without consuming energy warrant greater attention and wider application.