Oecologia (1992) 91:23-29
Oecologia 9 Springer-Verlag 1992
Earthworm diet related to soil organic matter dynamics through 13C measurements Agn6s Martin 1, J6rSme Balesdent 2' 3, and Andr6 Mariotti 2 1 Laboratoire d'Ecologie (CNRS UA 258), Ecole Normale Sup6rieure, 46 rue d'Ulm, F-75230 Paris Cedex 05, France 2 Laboratoire de Biog6ochimieIsotopique (CNRS UA 196-INRA) Universit6 Paris VI, 4 Place Jussieu, F-75252 Paris Cedex 05, France 3 Station de Science du Sol, INRA, Route de Saint-Cyr, F-78026 Versailles Cedex, France Received October 16, 1991 / Accepted in revised form February 27, 1992
Summary. The ability of earthworms to assimilate soil organic pools of different ages was investigated in field conditions through natural t3C labelling. Provided that ~3C natural abundance of earthworm tissues is determined by their diet, the assimilation of soil organic matter by earthworms was estimated by measuring i3 C/12 C ratio of tissues of earthworms sampled in soils containing organic pools with differing 13C/~2C ratios. Earthworms were collected in two locations where the vegetation shifted from a C3 to a C4 type (France), or from a C4 to a C3 type (Ivory Coast) 3 to 20 years ago. The results show that irrespective of their ecology (litter or soil-feeding), earthworms mainly feed on recent soil organic pools and assimilate soil organic matter with the same age distribution as the overall decomposers in the soil. Key words: Soil organic matter dynamics - Earthworm - Diet - 13C labelling
In many ecosystems, the soil fauna, and more particularly earthworms, is expected to play a determinant role on soil fertility (Lavelle et al. 1992) through their influence on water regime, aggregation, incorporation of litter, decomposition of soil organic matter (SOM) and microbial activity. The importance of earthworms on SOM decomposition and nutrient recycling needs to be clearly described in order to incorporate this soil fauna component into simulation models of SOM dynamics (see e.g. Parton et al. 1987; Jenkinson 1990). A major difficulty in evaluating the importance of earthworms in the soil carbon budget, is to identify the precise soil organic pools assimilated by the animals. Analyses of gut contents (Bouch6 and Kretzschmar 1974; Piearce 1978, Ferrirre 1980) or enzymatic equipment (Van Gansen 1962; Loquet and Vinceslas 1987) allow examination of what is potentially assimilable by the worm. Only labelling techniques can reveal which Correspondence to:
J. Balesdent
organic resources are really assimilated by the animals. However, labelling techniques that are currently used (14C or 15N) provide accurate tools only for testing the assimilation of young organic residues (Bouch6 1984; Cortez 1989). In this paper, we develop field labelling techniques based on t3C to identify the functional pools of SOM that are assimilated by earthworms, belonging to different ecological categories. Provided that the natural abundance of ~3C in earthworms tissue is determined by their diet (De Niro and Epstein 1978; Martin et al. 1992), assimilation of soil organic pools can be investigated by measuring gt3C ratio of the tissues of earthworms living in soils containing organic pools with distinct ~3C natural abundance. Earthworms were collected in three sites where a C3 to C4 or C4 to C3 vegetation shift had occured (temperate cropping system and tropical woodland). The ~3C labelling of earthworms resulting from the natural labelling of SOM induced by the shift of vegetation was used to test the ability of earthworms to assimilate soil organic pools of different ages. The isotopic composition of the animals was compared to the isotopic composition of the CO2 fluxes released by decomposition in a simple model of SOM dynamics to compare earthworms to other soil microbial decomposers.
Material and methods Experimental
sites
This study was performed in two regions each having distinct climatic regimes: Boignevilleand Versailles. The climate is defined as oceanic temperate (mean annual temperature: 10.5~ C, mean annual rainfall: 600 ram). Soils are eutric cambisols (FAO). The main characteristics of the topsoil are C=0.95 and 0.92%; clay=22.0 and 15.5%; C/N=8.5 and 9.1; pH=6.5 and 6.7 at Boigneville and Versailles, respectively. The soils are very similar and will be considered as the same throughout this study. Both sites are experimental fields designed for studies of longterm soil organic matter dynamics. At Boigneville, some plots has France - Ile de France.
24 been continuously cropped with C3 plants (C3 crops rotation, or C3 grasses) since deforestation, while in adjacent plots, vegetation shifted from C3 crops to continuous maize (C4 plant) in 1970. At Versailles, a plot previously cropped with C3 plants has been shifted similarly to maize in 1987. The natural labelling of SOM by 13C resulting from both shifts of vegetation is currently being studied and preliminary data have been reported in Balesdent et al. (1990) and Balabane and Balesdent (1992).
standard (Craig 1957) 1. Analytical precision on perfectly homogenized samples is a: 0.05%0. 613C measurements of earthworms were made on 3 or 4 different individuals; mean values and standard-errors were calculated. Statistical analyses consisted of A N O V A tests. The level of significance was defined as 0.05.
Calculation Ivory Coast - Baould region. Lamto. The climate is characterized by a high mean annual temperature (28 ~ C) and irregular annual rainfall (1300 mm with a dry season from November to March and in August). Soils are mainly Ferralsol (FAO); they are very sandy, low acidic and have low organic matter contents (5 to 12 m g C - g - i soil). In Lamto region, grass and shrub savannas represent the most extensive type of vegetation, due to the annual burning, and natural forests are generally restricted to river borders. Protection of savannas from fire results rapidely into recolonisation by trees; this phenomenon occured in a plot of grass savanna (C4 grasses) anthropieally protected from fire since 1963, and colonized by trees (C3 plants) since 1971. The natural labelling of the SOM by 13C resulting from this shifting of vegetation types has been measured in 1987 and corresponding data have been reported in Martin et al. (1990).
E a r t h w o r m samplin9 At each site, earthworms were sampled in two types of plots: plots in which a shift of vegetation type occured a few years ago (maize cropping systems at Boigneville and Versailles, savanna protected from fire at Lamto (see previous description)) were sampled to compare the C composition of SOM to that of earthworm tissues. - reference plots covered with unchanged type of vegetation (C3 gallery forest and C4 grass savanna at Lamto, continuous C3 cropping system and C3 grassland at Boigneville; no plot with continuous C4 vegetation are currently available in France) were sampled to provide reference values for earthworm tissues (see calculation section). In each plot, earthworms were collected at least 20 m away from the edges of the plots, in February 1990 at Versailles and Boigneville, and in August 1990 at Lamto. A complementary sampling was conducted in February 1991 at Versailles. Earthworms were separated according to ecological categories (Bouch6 1977) into: totally or partially pigmented species feeding on litter (litter-feeding earthworms), unpigmented species feeding on bulk soil (soil-feeding earth-
The distribution of C derived from both types of vegetation (C3 and C4 plants) in earthworm tissues was directly derived from the equation used to calculate the distribution of C3-C and C4-C derived C : (1)
6 = F62+(1-F)61, that gives: F -
6--6 i
(2)
62_61 ' where F is the percentage of C derived from the second type of vegetation (veg2) in earthworm tissues, 6 is the 613 C ratio of earthworms collected after the vegetation change, 61 and 62 are the 613 C ratios of earthworm tissues derived from vegetation 1 and vegetation 2, respectively. The exact values of 61 and 62 cannot be measured directly because the 613C of vegetation can be slightly changed during carbon decay in soil and earthworm metabolism (see discussion section). However, we assumed that 61 and 62 can be approximated with the 613C ratios measured for earthworms living in soil with permanent vegetation 1 or vegetation 2, respectively. Such an approximation, raised a major difficulty for earthworms sampled in french sites, as no soil under permanent C4 vegetation does currently exist in France. As a result, we calculated a theoritical 813C ratio for french earthworms living in an hypothetical soil under permanent C4 vegetation, assuming that the difference of 613C ratio between earthworms living in soil under permanent C3 vegetation and earthworms living in soil under permanent C4 vegetation is similar to the difference of 613C ratio between C3 vegetation and C4 vegetation: 62--61
=
(3)
6veg2--6vegi,
that gives: F' -
6--61
(4)
6veg2--6veg 1 '
-
WOrlTIS).
After taxonomic identification, earthworms were killed by immersion for ls in boiling water and then their fresh weight was determined. Gut content was removed, and the bodies were freezedried.
A n a l y t i c a l methodology The I3C natural abundance of earthworms was determined on the CO2 gas obtained after total combustion of organic matter in a sealed quartz tube with CuO at 900 ~ C. After breaking the tube in vacuo, the CO2 evolved was purified over pure reduced CuO heated at 600 ~ C, and dried and analyzed in an isotope ratio mass spectrometer (Finnigan Delta E and VG Sira 10) fitted with triple ion collector and dual inlet system equipped for rapid switching between reference and sample (Wedeking et al. 1983). Results are obtained in 613C%0 units. Laboratory references were calibrated using NSB19; results are expressed versus the international PDB
Both calculations of F and F' include an incertitude resulting from the assumption that, for each sampling site, the reference values of earthworms 613C ratios (rx and 62) are similar for adjacent plots; the uncertitude on earthworm reference values, that may reach up to 1 6 unit, leads to an incertitude on F and F', which may reach up to 7%. Provided the rapid turnover rate of earthworm tissues (less than 3 months) (Martin et al. 1992), we assumed that the 613C of earthworm reflects isotopic composition of organic components currently assimilated by the animals. As a result, the C3-C and C4-C distribution in earthworms tissues is assumed to reflect the C3 C and C4-C distribution in SOM assimilated by the animals.
/ 1 ~13 C = (
\
\ 13Rsample 13 Rstandar d
1] - 1000,
J
where
13C 13R = - 12 C
25 Table 1. Main characteristics of the surface layer of the soils ( ~ 1 0 cm at Lamto (Ivory Coast), 0-30 cm at Versailles and Boigneville (France)) where earthworms were collected (from Martin et al. (1990) and Balesdent et al. (1990 and 1992)). n a = n o n
available. Year of soil sampling: (1)1987, (2)1990, (3)1991, (4)1987. * : percentage of soil C derived from the second vegetation type; see calculation in the text pH
Ivory eoast Lamto
Shrub savanna (mainly C4 plants) 6.8 Savanna (C4 plants)+ 16 years woodland (C3 plants) (1) 6.8 Secondary Forest (C3 plants) 7.2
France Versailles
Continuous C3 crops C3 crops_+3 years maize (C4 plant) (z) C3 crops+4 years maize (C4 plant)(3)
France Boigneville
Continuous C3 crops C3 crops_+ 17 years maize (C4 plant) (4)
Table 2. 513C values and percentage of C3-C of earthworms collected at Lamto (Ivory Coast) in the 0-10 cm layer of a shrub savanna (C4 vegetation), a secondary forest soil (C3 vegetation), a savanna (C4 vegetation) shifted to a woodland (C3 vegetation) for 19 years.
- 11.8 _+0.4 a -23.8_+0.7 d -23.9_+0.5 d 101 _+3
Table 3.513 C values and percentage of C4~C of earthworms collected at Versailles and Boigneville (France) in the 0-10 cm layer of a grassland or a continuous C3 cropping system, and three C3 cropping systems cultivated with maize (C4 plant) for 3, 4 or 20 years
Plant (6ve~)
F of SOM* (%)
8.2 9.1 12.6
-12.7_+0.5 -22.0_+0.9 -26.6_+0.9
0 63_+7 100
6.5 6.5 6.5
9.2 9.2 9.2
-26.4_+0.2 -25.3_+0.1 --25.1_+0.1
0 8_+l 9_+1
6.7 6.5
11.2 9.5
--25.2_+0.1 --20.8_+0.2
0 24_+2
Eudrilidae Soil-feeder
D. agilis
M. lamtoiana
Litter-feeder
Litter-feeder
na --24.5_+0.4 d --23.l -+0.2 e 89 -+2**
na -26.4+0.8 c --25.5-+0.2 c 94 _+1"*
Soil-feeder 513C (%0) 613C ( % 0 ) 513C ( % o ) F(C3-C in %)*
513C of soil (%o)
Number of replicates = 4; the dispersion is given by standard deviation. Data with different letters are significantly different (P < 0.05). * see F definition in the text. ** Calculation made with 61 = - 11.8%o (M. anomala value)
Millsonia anomala Shrub savanna Secondary Forest Savanna+ 19 years woodland
C content m g C . g-1 soil
na -26.2+0.3 c --25.8-+0.9 c 96 _+5**
Calculation values 51 62 6
as compared to the plant value. The dispersion is given by the standard deviation. Number of replicates = 3 or 4 (1 if no standard deviation). Data with different letters are significantly different (P
Oecologia 9 Springer-Verlag 1992
Earthworm diet related to soil organic matter dynamics through 13C measurements Agn6s Martin 1, J6rSme Balesdent 2' 3, and Andr6 Mariotti 2 1 Laboratoire d'Ecologie (CNRS UA 258), Ecole Normale Sup6rieure, 46 rue d'Ulm, F-75230 Paris Cedex 05, France 2 Laboratoire de Biog6ochimieIsotopique (CNRS UA 196-INRA) Universit6 Paris VI, 4 Place Jussieu, F-75252 Paris Cedex 05, France 3 Station de Science du Sol, INRA, Route de Saint-Cyr, F-78026 Versailles Cedex, France Received October 16, 1991 / Accepted in revised form February 27, 1992
Summary. The ability of earthworms to assimilate soil organic pools of different ages was investigated in field conditions through natural t3C labelling. Provided that ~3C natural abundance of earthworm tissues is determined by their diet, the assimilation of soil organic matter by earthworms was estimated by measuring i3 C/12 C ratio of tissues of earthworms sampled in soils containing organic pools with differing 13C/~2C ratios. Earthworms were collected in two locations where the vegetation shifted from a C3 to a C4 type (France), or from a C4 to a C3 type (Ivory Coast) 3 to 20 years ago. The results show that irrespective of their ecology (litter or soil-feeding), earthworms mainly feed on recent soil organic pools and assimilate soil organic matter with the same age distribution as the overall decomposers in the soil. Key words: Soil organic matter dynamics - Earthworm - Diet - 13C labelling
In many ecosystems, the soil fauna, and more particularly earthworms, is expected to play a determinant role on soil fertility (Lavelle et al. 1992) through their influence on water regime, aggregation, incorporation of litter, decomposition of soil organic matter (SOM) and microbial activity. The importance of earthworms on SOM decomposition and nutrient recycling needs to be clearly described in order to incorporate this soil fauna component into simulation models of SOM dynamics (see e.g. Parton et al. 1987; Jenkinson 1990). A major difficulty in evaluating the importance of earthworms in the soil carbon budget, is to identify the precise soil organic pools assimilated by the animals. Analyses of gut contents (Bouch6 and Kretzschmar 1974; Piearce 1978, Ferrirre 1980) or enzymatic equipment (Van Gansen 1962; Loquet and Vinceslas 1987) allow examination of what is potentially assimilable by the worm. Only labelling techniques can reveal which Correspondence to:
J. Balesdent
organic resources are really assimilated by the animals. However, labelling techniques that are currently used (14C or 15N) provide accurate tools only for testing the assimilation of young organic residues (Bouch6 1984; Cortez 1989). In this paper, we develop field labelling techniques based on t3C to identify the functional pools of SOM that are assimilated by earthworms, belonging to different ecological categories. Provided that the natural abundance of ~3C in earthworms tissue is determined by their diet (De Niro and Epstein 1978; Martin et al. 1992), assimilation of soil organic pools can be investigated by measuring gt3C ratio of the tissues of earthworms living in soils containing organic pools with distinct ~3C natural abundance. Earthworms were collected in three sites where a C3 to C4 or C4 to C3 vegetation shift had occured (temperate cropping system and tropical woodland). The ~3C labelling of earthworms resulting from the natural labelling of SOM induced by the shift of vegetation was used to test the ability of earthworms to assimilate soil organic pools of different ages. The isotopic composition of the animals was compared to the isotopic composition of the CO2 fluxes released by decomposition in a simple model of SOM dynamics to compare earthworms to other soil microbial decomposers.
Material and methods Experimental
sites
This study was performed in two regions each having distinct climatic regimes: Boignevilleand Versailles. The climate is defined as oceanic temperate (mean annual temperature: 10.5~ C, mean annual rainfall: 600 ram). Soils are eutric cambisols (FAO). The main characteristics of the topsoil are C=0.95 and 0.92%; clay=22.0 and 15.5%; C/N=8.5 and 9.1; pH=6.5 and 6.7 at Boigneville and Versailles, respectively. The soils are very similar and will be considered as the same throughout this study. Both sites are experimental fields designed for studies of longterm soil organic matter dynamics. At Boigneville, some plots has France - Ile de France.
24 been continuously cropped with C3 plants (C3 crops rotation, or C3 grasses) since deforestation, while in adjacent plots, vegetation shifted from C3 crops to continuous maize (C4 plant) in 1970. At Versailles, a plot previously cropped with C3 plants has been shifted similarly to maize in 1987. The natural labelling of SOM by 13C resulting from both shifts of vegetation is currently being studied and preliminary data have been reported in Balesdent et al. (1990) and Balabane and Balesdent (1992).
standard (Craig 1957) 1. Analytical precision on perfectly homogenized samples is a: 0.05%0. 613C measurements of earthworms were made on 3 or 4 different individuals; mean values and standard-errors were calculated. Statistical analyses consisted of A N O V A tests. The level of significance was defined as 0.05.
Calculation Ivory Coast - Baould region. Lamto. The climate is characterized by a high mean annual temperature (28 ~ C) and irregular annual rainfall (1300 mm with a dry season from November to March and in August). Soils are mainly Ferralsol (FAO); they are very sandy, low acidic and have low organic matter contents (5 to 12 m g C - g - i soil). In Lamto region, grass and shrub savannas represent the most extensive type of vegetation, due to the annual burning, and natural forests are generally restricted to river borders. Protection of savannas from fire results rapidely into recolonisation by trees; this phenomenon occured in a plot of grass savanna (C4 grasses) anthropieally protected from fire since 1963, and colonized by trees (C3 plants) since 1971. The natural labelling of the SOM by 13C resulting from this shifting of vegetation types has been measured in 1987 and corresponding data have been reported in Martin et al. (1990).
E a r t h w o r m samplin9 At each site, earthworms were sampled in two types of plots: plots in which a shift of vegetation type occured a few years ago (maize cropping systems at Boigneville and Versailles, savanna protected from fire at Lamto (see previous description)) were sampled to compare the C composition of SOM to that of earthworm tissues. - reference plots covered with unchanged type of vegetation (C3 gallery forest and C4 grass savanna at Lamto, continuous C3 cropping system and C3 grassland at Boigneville; no plot with continuous C4 vegetation are currently available in France) were sampled to provide reference values for earthworm tissues (see calculation section). In each plot, earthworms were collected at least 20 m away from the edges of the plots, in February 1990 at Versailles and Boigneville, and in August 1990 at Lamto. A complementary sampling was conducted in February 1991 at Versailles. Earthworms were separated according to ecological categories (Bouch6 1977) into: totally or partially pigmented species feeding on litter (litter-feeding earthworms), unpigmented species feeding on bulk soil (soil-feeding earth-
The distribution of C derived from both types of vegetation (C3 and C4 plants) in earthworm tissues was directly derived from the equation used to calculate the distribution of C3-C and C4-C derived C : (1)
6 = F62+(1-F)61, that gives: F -
6--6 i
(2)
62_61 ' where F is the percentage of C derived from the second type of vegetation (veg2) in earthworm tissues, 6 is the 613 C ratio of earthworms collected after the vegetation change, 61 and 62 are the 613 C ratios of earthworm tissues derived from vegetation 1 and vegetation 2, respectively. The exact values of 61 and 62 cannot be measured directly because the 613C of vegetation can be slightly changed during carbon decay in soil and earthworm metabolism (see discussion section). However, we assumed that 61 and 62 can be approximated with the 613C ratios measured for earthworms living in soil with permanent vegetation 1 or vegetation 2, respectively. Such an approximation, raised a major difficulty for earthworms sampled in french sites, as no soil under permanent C4 vegetation does currently exist in France. As a result, we calculated a theoritical 813C ratio for french earthworms living in an hypothetical soil under permanent C4 vegetation, assuming that the difference of 613C ratio between earthworms living in soil under permanent C3 vegetation and earthworms living in soil under permanent C4 vegetation is similar to the difference of 613C ratio between C3 vegetation and C4 vegetation: 62--61
=
(3)
6veg2--6vegi,
that gives: F' -
6--61
(4)
6veg2--6veg 1 '
-
WOrlTIS).
After taxonomic identification, earthworms were killed by immersion for ls in boiling water and then their fresh weight was determined. Gut content was removed, and the bodies were freezedried.
A n a l y t i c a l methodology The I3C natural abundance of earthworms was determined on the CO2 gas obtained after total combustion of organic matter in a sealed quartz tube with CuO at 900 ~ C. After breaking the tube in vacuo, the CO2 evolved was purified over pure reduced CuO heated at 600 ~ C, and dried and analyzed in an isotope ratio mass spectrometer (Finnigan Delta E and VG Sira 10) fitted with triple ion collector and dual inlet system equipped for rapid switching between reference and sample (Wedeking et al. 1983). Results are obtained in 613C%0 units. Laboratory references were calibrated using NSB19; results are expressed versus the international PDB
Both calculations of F and F' include an incertitude resulting from the assumption that, for each sampling site, the reference values of earthworms 613C ratios (rx and 62) are similar for adjacent plots; the uncertitude on earthworm reference values, that may reach up to 1 6 unit, leads to an incertitude on F and F', which may reach up to 7%. Provided the rapid turnover rate of earthworm tissues (less than 3 months) (Martin et al. 1992), we assumed that the 613C of earthworm reflects isotopic composition of organic components currently assimilated by the animals. As a result, the C3-C and C4-C distribution in earthworms tissues is assumed to reflect the C3 C and C4-C distribution in SOM assimilated by the animals.
/ 1 ~13 C = (
\
\ 13Rsample 13 Rstandar d
1] - 1000,
J
where
13C 13R = - 12 C
25 Table 1. Main characteristics of the surface layer of the soils ( ~ 1 0 cm at Lamto (Ivory Coast), 0-30 cm at Versailles and Boigneville (France)) where earthworms were collected (from Martin et al. (1990) and Balesdent et al. (1990 and 1992)). n a = n o n
available. Year of soil sampling: (1)1987, (2)1990, (3)1991, (4)1987. * : percentage of soil C derived from the second vegetation type; see calculation in the text pH
Ivory eoast Lamto
Shrub savanna (mainly C4 plants) 6.8 Savanna (C4 plants)+ 16 years woodland (C3 plants) (1) 6.8 Secondary Forest (C3 plants) 7.2
France Versailles
Continuous C3 crops C3 crops_+3 years maize (C4 plant) (z) C3 crops+4 years maize (C4 plant)(3)
France Boigneville
Continuous C3 crops C3 crops_+ 17 years maize (C4 plant) (4)
Table 2. 513C values and percentage of C3-C of earthworms collected at Lamto (Ivory Coast) in the 0-10 cm layer of a shrub savanna (C4 vegetation), a secondary forest soil (C3 vegetation), a savanna (C4 vegetation) shifted to a woodland (C3 vegetation) for 19 years.
- 11.8 _+0.4 a -23.8_+0.7 d -23.9_+0.5 d 101 _+3
Table 3.513 C values and percentage of C4~C of earthworms collected at Versailles and Boigneville (France) in the 0-10 cm layer of a grassland or a continuous C3 cropping system, and three C3 cropping systems cultivated with maize (C4 plant) for 3, 4 or 20 years
Plant (6ve~)
F of SOM* (%)
8.2 9.1 12.6
-12.7_+0.5 -22.0_+0.9 -26.6_+0.9
0 63_+7 100
6.5 6.5 6.5
9.2 9.2 9.2
-26.4_+0.2 -25.3_+0.1 --25.1_+0.1
0 8_+l 9_+1
6.7 6.5
11.2 9.5
--25.2_+0.1 --20.8_+0.2
0 24_+2
Eudrilidae Soil-feeder
D. agilis
M. lamtoiana
Litter-feeder
Litter-feeder
na --24.5_+0.4 d --23.l -+0.2 e 89 -+2**
na -26.4+0.8 c --25.5-+0.2 c 94 _+1"*
Soil-feeder 513C (%0) 613C ( % 0 ) 513C ( % o ) F(C3-C in %)*
513C of soil (%o)
Number of replicates = 4; the dispersion is given by standard deviation. Data with different letters are significantly different (P < 0.05). * see F definition in the text. ** Calculation made with 61 = - 11.8%o (M. anomala value)
Millsonia anomala Shrub savanna Secondary Forest Savanna+ 19 years woodland
C content m g C . g-1 soil
na -26.2+0.3 c --25.8-+0.9 c 96 _+5**
Calculation values 51 62 6
as compared to the plant value. The dispersion is given by the standard deviation. Number of replicates = 3 or 4 (1 if no standard deviation). Data with different letters are significantly different (P
