Morphology

A plant fiber is an elongated, narrow cell, with sharp ends; it has a lumen or cavity and a wall constituted, in turn, by two walls, an external one, called primary wall, and an internal one, called secondary wall ^{Azcárraga et al., 2010, p. 93}); it can present cross-linking markings and dislocations or nodes. Such marks are striations transverse to the fiber’s surface-sometimes also presenting longitudinal striations-so much so that it is possible to observe their faint or strong markings (^{Catling & Grayson, 2004, p. 3}; ^{Suomela et al., 2017, p. 413}). Dislocations are visible as ring-shaped marks, coarsening or knots that are generated by a change in fiber direction due to external or internal stresses (^{Catling & Grayson, 2004, p. 1}; ^{Suomela et al., 2017, p. 413}) (Figure 4). It should be added that there is no consensus in literature on the origin and definition of these markings.

FIGURE 4 Photomicrographs of standard fibers: flax and hemp 

Pattern	Section type		Microchemical with Schweitzer’s reagent	Modified Herzog test
	Longitudinal	Cross
	A	B	C	D
Linen
	Fibers with “X” X-shaped dislocations and crossover marks	Thick-walled, four-, five, and six-sided fibers with flat surfaces with reduced lumens	Spiral fiber contraction	Color change at 90° yellow and at 0° blue
Hemp
	Fibers with well-marked dislocations and some crossing marks	Thick-walled fibers and small lumens	Fiber shrinkage in folds	Color change at 90° yellow and at 0° blue

Source: (Optical microscopy and microchemistry: Victor Santos Vázquez, 2020; micrographs of modified Herzog’s test: Ana Laura Avelar-Carmona, 2020; courtesy: Laboratorio de Biología, Escuela Nacional de Conservación, Restauración y Museografía [ENCRyM] and Instituto Nacional de Antropología e Historia [INAH]).

Modified Herzog test or delayed filter test (λ)

This analysis is based on the fact that plant fibers are composed of a lumen or cavity and a cell wall containing various chemical compounds, mainly cellulose, and, among others, lignin and hemicellulose: for example, flax contains 71% cellulose and 2% lignin, and hemp, 75% and 4% respectively (^{Avérous & Le Digabel, 2006}). The cell wall, in turn, is made up of a primary and a secondary wall (three different layers make up the latter), both of which are formed by microfibrils, and these in turn by more microfibrils-cellulose chains with birefringent³ properties; those of the thickest layer of the secondary wall have a helical arrangement and can rotate either to the right (forming S helices) or to the left (forming Z helices), determining the microfibrillar orientation according to the letters S or Z, which is the path taken by the microfibrils with respect to the observer (^{Strasburger et al., 2004, p. 95}). Flax has a characteristic S twist, while hemp has a Z twist (^{Strasburger et al., 2004, p. 95}; ^{Haugan & Holst, 2013, p. 160}; ^{Suomela et al., 2017, p. 414}), varying the birefringent properties of each.

This property allows the Herzog test to be applied using a microscope with a polarizer. If the type of fiber present has a Z-twist, as in the case of hemp, when parallel to the polarizer (vertical) it will turn blue, when parallel to the analyzer (horizontal), yellow, and then the opposite will be true: yellow vertically and blue horizontally, when the fiber has an S-twist, as is characteristic of flax (Figure 4. Flax D and Hemp D) (^{Haugan & Holst, 2013, pp. 166-167}). For the accurate execution of this test, it is necessary to first set the microscope parameters using reference fibers (Suomela et al., 2018, p. 418) (Figure 4).

Microchemical analysis: Schweitzer’s reagent test

This analysis technique is specific for vegetable fibers and has a destructive effect, so it is only recommended to be used as an alternative to optical microscopy, if a sufficient amount of the sample is available. It is useful to differentiate between linen and hemp fibers, mainly when the state of deterioration of the specimens prevents observing the morphological features of the fibers by means of optical microscopy.

In this test the fiber is put into contact with an ammoniacal solution of cupric hydroxide capable of dissolving the cellulose, after which a cellulose-copper complex is formed. Once the cell wall is dissolved, the resulting filament acquires feature shapes depending on the cellulose distribution. When a flax fiber is subjected to this test, the lumen contracts forming a spiral, whereas, in hemp, the internal cavity forms transverse folds (^{Campo et al., 2009, p. 25}) (Figure 4. Flax C and Hemp A). Here we follow the methodology of ^{T. A. Geissman (1974)} applied effectively on heritage objects (^{Juanes et al., 2008, p. 75}). Figure 4 shows the microphotographs of two standard samples analyzed by the three methods described.

A. Optical microscopy (OM) with polarized light (PLM) and transmitted light (TL).

Morphological analysis of the fibers (length, walls, dislocations, cross-linking marks, lumen and general shape) was performed with an Olympus BX51 optical microscope equipped with an Olympus DP72 high-resolution camera. The images were processed using Olympus Cell-Sens Dimension software^®. The aforementioned equipment is located in the facilities of the Chemistry-Biology Laboratory of the National Coordination for the Conservation of Cultural Heritage (CNCPC) of the National Institute of Anthropology and History (INAH).

B. Modified Herzog test

This analysis was performed based on the Herzog test, modified by ^{Haugan and Holst (2013, p. 166)}, for which an Olympus BX43 petrographic microscope with an Olympus EG20 integrated camera was used. The images were processed using AxioVision software.

To conduct this test it is necessary, to first perform a search using the 10X lens the fiber on which the analysis will be carried out. Once located, it must be positioned, with the help of the rotating plate, at an angle α = 0°. Subsequently, the objective is changed to 40X: the fiber is then observed with crossed polars/nicols and the analyzer is set until the fiber is brought to the extinction angle, which is achieved when it, or a fragment of it, is extinguished, i.e., it “disappears” or darkens.

Then, the compensator or λ-foil is inserted, at which time the background should appear magenta or pink, then focusing our attention on a small segment of the fiber. Next, we rotate the stage at an angle of α = 90° and note the color change of the same section observed before α = 0°.

C. Preparation of Schweitzer’s reagent

The reagent for identification is prepared by dissolving copper sulfate in hot water. Then a 28% ammonia solution is added drop by drop, until the pale blue color of the copper sulfate is diminished, and the solution becomes transparent. This solution is then rinsed with water, filtered, and from the precipitate two or three drops of ammonium hydroxide are added, which gives an intense dark blue color. It is important to mention that this reagent is functional when freshly prepared before crystallization; after that occurs, it loses its effectiveness. Once the drop of the reagent is applied to the fiber, the shrinkage of the material is observed under optical microscopy (OM) and using polarized light (PLM), the images are captured.

Discussion of results

Figure 6 summarizes the results of the three complementary techniques used in this research for the precise identification of Liberian fibers. The yellow color highlights the properties assigned to flax according to the literature consulted, and in green, those related to hemp. The above shows that most of the analyzed fibers observed characteristics -especially- morphological characteristics of both flax and hemp, so that the correct interpretation of the constituent material of the analyzed paints was conducted by comparing the results obtained from both the modified Herzog test and with the chemical reaction of Schweitzer.

Figures 7 and 8 show the evidence of the results obtained in each of the tests performed on the samples of the fibers of the support of the New Spanish paintings.

FIGURE 7 Morphology  

Painting	Longitudinal section		Cross section
	A	B	C
NE235 M1
NE229C M5
NE229B M5
NE245C M6
NE233 M4
NE232 M4
NE245CM7
NE233 M5
NE232 M5

Source: (Micrographs: Victor Santos Vázquez, 2020; courtesy: Laboratorio de Química, Research area of the CNCPC-INAH).

FIGURE 8 Results of microchemical tests and the Herzog test 

Sample	Microchemistry Schweitzer’s reagent	Herzog test
	A	B
NE235M1
NE229C M5
NE229B M5
NE245C M6
NE233 M4
NE232 M4
NE245C M7
NE233 M5	Pliegue
NE232 M5

Source: (Optical microscopy and microchemistry: Victor Santos Vázquez, 2020; micrographs of modified Herzog’s test: Ana Laura Avelar-Carmona, 2020; courtesy: Laboratorio de Biología, Escuela Nacional de Conservación, Restauración y Museografía [ENCRyM], Instituto Nacional de Antropología e Historia [INAH]).

All the samples of Liberian fibers analyzed presented dislocations or X-shaped nodes; therefore, this morphological feature does not allow a precise identification of the material used as support in the paintings (Figure 7. A and B).

Of the nine samples studied, only three: NE229BM5, NE232M4 and NE233M4, presented in their entirety morphological characteristics reported in the known properties of flax, that is, equidistantly distributed dislocations and reduced and discontinuous lumen that closes in sections (^{Campo et al., 2009: 18}; ^{García, 2007, p. 137}; Florian et al., 1994, p. 49; ^{Hall & Davies, 1968, p. 23}; ^{Mattews, 1934, p. 750}). Additionally, they came up positive for flax in the modified Herzog and microchemical tests.

Samples NE235M1 and NE232M5 presented dislocations without any specific distribution, which is indicative of hemp fibers (^{Campo et al., 2004, p. 19}; ^{Catling & Grayson, 2004, p. 22}; ^{García, 2007, p. 134}); however, their lumen went from continuous to discontinuous and reduced, as it is supposed to occur with flax fibers. Finally, these samples tested positive for hemp with Schweitzer’s reagent and the modified Herzog’s test, highlighting the importance of using a combined methodology for their correct identification.

In fact, samples NE229CM5, NE233M5, NE235M1, NE232M5, NE245CM6 and NE245CM7, i.e., more than half of the fibers under study, were representative examples of the problems faced by a researcher during the process of analyzing heritage materials. Under the optical microscope they presented morphological elements corresponding to flax, hemp or even both fibers. Only by applying the complementary tests of Herzog and microchemistry with Schweitzer’s reagent were results obtained that helped to make a correct identification. For example, sample NE245CM7 presented X-shaped dislocations at regular distances, as would flax, and its lumen was wide and continuous, a characteristic associated with hemp. The microchemical test was inconclusive, as the fiber contracted in an unidentified form, and only the Herzog test was positive for flax.

The shape of the fibers in cross section did not present an important value for their identification, since both hemp and flax fibers observed triangular, square, pentagonal, hexagonal, straight and rounded edges (Figure 7. C), which corresponds with ^{Lukesova & Holst (2020, p. 9)}, who also found no significant differences between the cross-sections of the two organic materials.

By carrying out a complementary analysis methodology with microchemistry and the Herzog test, it was possible to establish a different identification of the resulting morphology. Only the combination of analyses can guarantee reliable results, especially considering that in ancient paintings both the state of conservation of the material and even the process of plant growth, extraction and the complete processing of textile production modify the physical characteristics of a fiber.

In the identification methodology we used, we considered it a necessary criterion that at least two analyses should match, and to validate the results further it was necessary to apply the tests to more than one fiber. Regarding the test performed with Schweitzer’s reagent, it rarely matched with the other analyses, due to the amount of fiber used, the speed of the reaction and the degree of subjectivity involved in the interpretation of the resulting contractions. In all the cases analyzed, the Herzog test also linked with the other techniques used and was considered decisive in achieving the identifications. Therefore, its effectiveness in differentiating between flax and hemp was confirmed, even considering the thickness and deformations that the fibers may undergo, either by deterioration, by the growth of the fibers themselves or due to the textile manufacturing process.