2.3.2. EC1:5 Method

For the 1:5 suspension, 50 g of soil and 250 mL of distilled water were used. Three alternative methods were applied: the method of Loveday [18], the NRCS [15] and the USDA [2].

In the Loveday method, the suspension was shaken by a mechanical shaker for exactly one hour and then kept at rest for 20 min. After the rest time, the extract was obtained, and the EC was determined. For the NRCS method, the suspension remains at rest in complete shade for 23 h and then shaken mechanically for one hour. After the shaking, the extract was obtained, and the EC was determined. Finally, in the USDA method the suspension was shaken by hand, 4 times, every half hour for 30 s. After, the extract was obtained, and the EC was determined. The method of vacuum filtration in all the three methods is the same and common, followed by the measurement of EC with a conductivity meter. All the methods and EC readings were conducted at 25 ◦C.

In two soil samples, one from Laconia (sample L) and one from Argolida (sample A) with EC<sup>e</sup> values of 0.793 and 13.78 dS m−<sup>1</sup> , respectively, the EC1:5 values were measured after the suspensions were agitated with mechanical shaker for times 1, 2, 3, 4, 6, 24 and 48 h. After each agitation time the extraction was obtained, and the EC was determined. This process can better evaluate the role of shaking time on the EC1:5 values for the two very different EC<sup>e</sup> values.

## 2.3.3. EC1:1 Method

In the 1:1 method, the three above mentioned methods (Loveday, NRCS and USDA) were also applied as described in the 1:5 method. For each of the above methods, 50 g of soil was weighed and then each procedure was performed in the same way as above.

#### 2.3.4. Statistical Analysis

For the relationships EC<sup>e</sup> = f(EC1:1) and EC<sup>e</sup> = f(EC1:5), a least-squared linear regression was applied and the coefficient of determination R<sup>2</sup> was evaluated. The R<sup>2</sup> coefficient is used to assessing the correlation between two independent methods. Also, the values of root mean square errors (RMSE) were determined. Analysis of variance (ANOVA) was applied to test the significant difference among the applied EC1:5 or EC1:1 methods using SPSS Statistical Software v. 17.0 (SPSS Inc., Chicago, IL, USA); the means of each method were compared using t-test at a probability level P = 0.05.

#### **3. Results and Discussion**

#### *3.1. Soil Properties*

Samples from Laconia and Argolida are characterized as clay-clay loam soils and from Kos as sandy clay soils. All soil samples presented negligible gypsum content. As regards to CaCO3, samples from Laconia presented a content lower than 2.5%, from Argolida 5–8% and from Kos 8.5–11%. The pH values ranged from 7.69 to 8.06 for soil samples from Laconia and from 7.5 to 7.7 for soil samples from Argolida and Kos.

Additionally, the soil texture analyses of the two soil samples examined separately resulted as follows: (i) soil sample L—clay soil (23.5% sand, 16% silt, 60.5% clay) and (ii) soil sample A—clay loam/loam soil (39% sand, 32% silt, 29% clay). The CaCO<sup>3</sup> content was 0.2% and 7.66% and pH values were 7.75 and 7 for sample L and A, respectively.

#### *3.2. Estimation of Soil Salinity*

The EC<sup>e</sup> values ranged from 0.611 to 25.9 dS m−<sup>1</sup> . It should also be noted that the EC<sup>e</sup> variation range of the soil samples from Laconia is much lower than that of the other two regions (Argolida and Kos). Specifically, EC<sup>e</sup> values of the samples from Laconia ranged from 0.611 to 1.664 dS m−<sup>1</sup> , while in the other two regions they ranged from 2.32 to 25.9 dS m−<sup>1</sup> . From the measured EC<sup>e</sup> values, it appears that a relatively wide range in salinity levels was obtained for both comparing the different EC1:5 and EC1:1 methods, as well as evaluating the relationship between the EC<sup>e</sup> and each of EC1:5 or EC1:1 methods.

As regards to SP all soil samples examined (with exception of the two separated samples) have values greater than 43%, percentage which indicates that the soils are classified in fine textured soils [20]. More specifically, SP values ranged from 50.5% to 72.5% for soils from Laconia, 52–70% for soils from Argolida and 43–53% for soils from Kos.

#### *3.3. Comparison of 1:1 and 1:5 Soil to Water Extract Electrical Conductivity Methods*

In Table 2 the slope of the linear relationship (y = ax) between 1:5 soil to water extract electrical conductivity methods for EC<sup>e</sup> < 3 dS m−<sup>1</sup> and EC<sup>e</sup> > 3 dS m−<sup>1</sup> and R<sup>2</sup> are presented.

**Table 2.** Slopes of the linear equations describing the relation between 1:5 soil to water extract electrical conductivity methods for EC<sup>e</sup> < 3 dS m−<sup>1</sup> and EC<sup>e</sup> > 3 dS m−<sup>1</sup> and coefficient of determination R<sup>2</sup> .


Similarly, the slope and R<sup>2</sup> of the linear relationship between 1:1 soil to water extract electrical conductivity methods for EC<sup>e</sup> < 3 dS m−<sup>1</sup> and EC<sup>e</sup> >3 dS m−<sup>1</sup> are presented in Table 3.

**Table 3.** Slopes of the linear equations describing the relation between 1:1 soil to water extract electrical conductivity methods for EC<sup>e</sup> < 3 dS m−<sup>1</sup> and EC<sup>e</sup> > 3 dS m−<sup>1</sup> and coefficient of determination R<sup>2</sup> .


From the results presented in Tables 2 and 3, it is obvious that each of the three methods examined resulted in different values of both EC1:1 and EC1:5 when EC<sup>e</sup> < 3 dS m−<sup>1</sup> . Analysis of variance (ANOVA) showed that the three methods are significantly different at a probability level P = 0.05. Furthermore, the t-test analysis (P = 0.05) showed that the NRCS and Loveday methods as well as the USDA and Loveday methods resulted in significantly different EC1:5 values, while EC1:5 values between NRCS and USDA were not significantly different. The mean value with standard deviation for NRCS, USDA and Loveday methods were 0.177 <sup>±</sup> 0.029, 0.169 <sup>±</sup> 0.029 and 0.151 <sup>±</sup> 0.027 dS m−<sup>1</sup> , respectively. In the case of 1:1 ratio, the EC values between NRCS and USDA as well as NRCS and Loveday methods were also significantly different (P = 0.05). The mean value with standard deviation for NRCS, USDA and Loveday methods were 0.5 <sup>±</sup> 0.070, 0.43 <sup>±</sup> 0.100 and 0.423 <sup>±</sup> 0.086 dS m−<sup>1</sup> , respectively.

The NRCS method resulted in greater EC values compared to the other two methods for both 1:1 and 1:5 ratios, whereas the Loveday method resulted in lower EC values. From these results, it appears that at low values of EC<sup>e</sup> (EC<sup>e</sup> < 3 dS m−<sup>1</sup> ) the rest time seems to play an important role since the difference between the NRCS and the Loveday method is only in the duration of rest time. As regards to the NRCS and USDA methods, the slope of the linear regression between the NRCS and USDA at 1:5 ratio is 1.047, while at 1:1 is 1.161.

The EC1:5 values of the soil sample L (with EC<sup>e</sup> = 0.793 dS m−<sup>1</sup> < 3 dS m−<sup>1</sup> ) obtained by mechanical shaking for 1, 2, 3, 4 and 6 h was approximately 0.142 dS m−<sup>1</sup> while EC1:5 values for 24 and 48 h were 0.218 and 0.274 dS m−<sup>1</sup> , respectively. Practically, after 48 h shaking the EC1:5 value was approximately doubling. The corresponding EC values obtained by the three methods used were 0.141, 0.127 and 0.158 dS m−<sup>1</sup> for USDA, Loveday and NRCS methods, respectively. Therefore, it appears that the agitation time plays a dominant role to obtain equilibrium since the difference between the NRCS method (EC1:5 = 0.158 dS m−<sup>1</sup> ) and the method with 24 h shaking (EC1:5 = 0.218 dS m−<sup>1</sup> ) is in the shaking time. These results are similar to those of He et al. [6] in terms of the long shaking time required to equilibration but differ in the fact that in our experiments did not show differences in EC values obtained by shaking of at least up to 6 h. He et al. [6] explained that the higher values of EC obtained by the long shaking time method compared to other methods may be due to the fact that the mechanical shaking destroys micro-aggregates, as well as increase dissolution of salts because the dynamic concentration gradient between solid and liquid phases. Also, Vanderheynst et al. [12] found that differences occur for shaking time greater than a threshold value of 3 h.

In the case of soils with EC<sup>e</sup> > 3 dS m−<sup>1</sup> there is no significant differences between agitation methods since all methods gave almost the same results and the slope of the linear relationship is almost 1 (Tables 2 and 3). In addition, it is noted that the R<sup>2</sup> values for soils with EC<sup>e</sup> > 3 dS m−<sup>1</sup> are higher for all methods examined, in both 1:5 and 1:1 ratios, compared to R<sup>2</sup> values for EC<sup>e</sup> < 3 dS m−<sup>1</sup> (Tables 2 and 3).

The EC1:5 values of the soil sample A (with EC<sup>e</sup> = 13.8 dS m−<sup>1</sup> > 3 dS m−<sup>1</sup> ) obtained by mechanical shaking for 1, 2, 3, 4, 6, 24 and 48 h ranged from 1.683 to 1.751 dS m−<sup>1</sup> . It is obvious that for soils with EC<sup>e</sup> > 3 dS m−<sup>1</sup> the shaking times required to obtain equilibration are significantly lower compared to soils with EC<sup>e</sup> < 3 dS m−<sup>1</sup>

The different behavior depending on the EC<sup>e</sup> value shows that the solid and liquid phases is far from considered a simple system where the only process carried out is dissolution and that the concentration of ions is inversely proportional to dilution. Such situations may exist only in sandy or sandy loam soils in semi-arid areas with high salinity [25]. However, the soils are characterized by a cation exchange capacity value depending on the type and quantity of clay, the presence of slightly soluble minerals but also ion exchanges between solid and liquid phase. In the present experimental work, the existence of a relatively high clay percentage combined with the existence of slightly soluble minerals may be led to different EC values among various methods, especially when EC<sup>e</sup> < 3 dS m−<sup>1</sup> . This phenomenon may be even more pronounced in the case of clay soils where there are high content of slightly soluble minerals but less pronounced in the coarse-textured soils without slightly soluble minerals.
