Radular force performance of stylommatophoran gastropods (Mollusca) with distinct body masses

The forces exerted by the animal’s food processing structures can be important parameters when studying trophic specializations to specific food spectra. Even though molluscs represent the second largest animal phylum, exhibiting an incredible biodiversity accompanied by the establishment of distinct ecological niches including the foraging on a variety of ingesta types, only few studies focused on the biomechanical performance of their feeding organs. To lay a keystone for future research in this direction, we investigated the in vivo forces exerted by the molluscan food gathering and processing structure, the radula, for five stylommatophoran species (Gastropoda). The chosen species and individuals have a similar radular morphology and motion, but as they represent different body mass classes, we were enabled to relate the forces to body mass. Radular forces were measured along two axes using force transducers which allowed us to correlate forces with the distinct phases of radular motion. A radular force quotient, AFQ = mean Absolute Force/bodymass0.67, of 4.3 could be determined which can be used further for the prediction of forces generated in Gastropoda. Additionally, some specimens were dissected and the radular musculature mass as well as the radular mass and dimensions were documented. Our results depict the positive correlation between body mass, radular musculature mass, and exerted force. Additionally, it was clearly observed that the radular motion phases, exerting the highest forces during feeding, changed with regard to the ingesta size: all smaller gastropods rather approached the food by a horizontal, sawing-like radular motion leading to the consumption of rather small food particles, whereas larger gastropods rather pulled the ingesta in vertical direction by radula and jaw resulting in the tearing of larger pieces.

The typical force exerted by feeding organs is a useful parameter indicating specializations to distinct food types, as it correlates with the food spectrum (see e.g. 1 ; for a review for stress-related puncture mechanics, see 2 ). This topic had been studied quite intensively in vertebrates (for a summary of the relevant literature, see 3 ): bite force analyses had been performed on mammals (e.g. [4][5][6][7][8], reptiles (e.g. [9][10][11][12], fish (e.g. [13][14][15] ), and birds (e.g. 16,17 ). Even though the majority of animal species belong to the invertebrates, unfortunately fewer work focused on the forces exerted by their structures involved in either gathering or acquiring food due to the difficulties of an experimental set-up for studies of small structures. Exceptions and pioneers in this field are studies performed on representatives of Arthropoda: spiders, crustaceans, scorpions, and insects [18][19][20][21][22][23][24] .
For molluscs, even though they represent the second specious animal group 25 with around 80,000 recent species only within the Gastropoda 26 , only a few studies approached the forces exerted by their feeding organ, the radula. Since the species belonging to this animal phylum occupy almost any marine, freshwater or terrestrial environment and established extremely varied ecological niches 27 , accompanied with feeding on a wide range of food sources with various mechanical properties, mollusc trophic specializations are of very high interest for evolutionary biologists. Their radula, one important molluscan autapomorphy and the interface between the organism and its ingesta (food, minerals), is highly diverse and offers an immense opportunity to study the structural adaptations enabling feeding on distinct ingesta types 28 .
The radula consists of a thin, chitinous membrane with rows of embedded, sometimes mineralized teeth which is supported by thick, underlying odontophoral cartilages and moved by numerous muscles of the buccal mass. The sometimes highly complex radular motion (see e.g. [29][30][31][32][33][34] ) brings the tooth cusps in contact with the ingesta leading to the tearing, cutting, and gathering of food.
The majority of previous studies have focused on the radular teeth themselves, but the forces exerted by this organ or its biomechanical performance have unfortunately only been investigated in a few papers devoted (1) to the feeding force calculations ( 83 ; or force calculations for radula-inspired gripping devices: 84 ), (2) to the experiments revealing the forces needed to remove algae 85,86 , or (3) to the first in vivo experiment performed on a single mollusc species 87 .
Here, before the broad topic of trophic specialization in molluscs can be approached and, to lay a keystone for further studies, we investigated the in vivo radular forces while foraging for five different stylommatophoran species using force transducers following the protocol of 87 . As the radular motion or radular morphology could potentially influence the forces exerted by the feeding organ we first selected species with a similar radular type (isodont) and similar radular motion to get a good impression on radular forces without being confronted with overflowing radular diversity, which could make results less comparable or prone to artefacts. Additionally, since the chosen gastropods however represent distinct body mass classes, we tested if and to which extend the chosen stylommatophoran specimens follow common laws and predictions for scaling of force and body mass.

Material and methods
For force measurements we have chosen five stylommatophoran species (Gastropoda: Heterobranchia) that were easy to obtain, as they are either often kept as pets or could be collected easily around Hamburg. Additionally, all these species possess a similar radular type with numerous, similar shaped (isodont), and small teeth, thus the direct influence of the tooth morphology on the forces produced is probably rather small. Also, stylommatophoran gastropods show a rather similar radular foraging motion, which can be described as a licking motion (for details on motion and radular type see e.g. 34,[87][88][89][90][91][92][93] ). There are surely differences in motion between species, because the arrangement or thickness of radular muscles might differ. Additionally, individual gastropods of the same species might also prefer a slightly different feeding motion, but the broad cycles of the pro-and retraction are comparable.
We have chosen the following gastropods, since they can be sorted to three distinct body mass classes (see Fig. 1 Supplementary Table 1).
Forces exerted by the radula were measured following the protocol of 87 . Snails were placed on an acrylic platform with a small hole of 4 mm diameter. The platform was attached to a laboratory jack so that the height could be adjusted. Food stripes (sliced to pieces of 3 [width] × 2 [length] × 20 or 40 [height] mm; either carrot (for all species except A. vulgaris) or fresh strawberry (for A. vulgaris), depending on the specific preference of the species, were glued to a needle, which was mounted onto a force transducer FORT-10 (World Precision Instruments, Sarasota, FL, USA) and stuck through the hole so that the snail could feed on it, but without the involvement of the foot (Fig. 2). A 1000 g sensor was used for mature L. fulica, for all other individuals a 25 g sensor was used. Forces could only be measured in either vertical or horizontal direction, but not in both simultaneously. Thus the experimental set-up was remodeled to receive data for both directions (see also 87 ). The force transducers were connected to an amplifier (Biopac System, Inc., CA, USA) and a computer-based data acquisition and processing system (AcqKnowledge™, Biopac Systems, Inc., v.3.7.0.0, World Precision Instruments, Sarasota, FL, www.nature.com/scientificreports/ USA). Not all force peaks were analyzed in AcqKnowledge™ due to the large sample size, but about 30 maximal and minimal force peaks were evaluated per feeding unit, which is the time needed for eating up the part of the carrot or strawberry that was not coated with glue. For the detailed quantity of evaluated force peaks per species, individual, and direction see Supplementary Detailed radular motion, while feeding on a flat surface was documented with a Keyence VHX-500 digital microscope (KEYENCE, Neu-Isenburg, Germany) by placing the individual on an acrylic platform, providing flour paste as food (see also 34,87 ). The behavior, while foraging on a carrot, was documented with an iPad Pro The shells of dead specimens were removed and the soft parts were weighed to receive body mass without shell. Forces were corrected for body mass without shell to receive Relative Force II (force/body mass without shell, mN/g). To estimate the mass of the entire buccal mass (BRJ), the radula and jaw (RJ), and the buccal mass musculature (B) these specimens were dissected which was documented with a Keyence VHX-500 digital microscope (KEYENCE, Neu-Isenburg, Germany). The BRJ was first extracted, freed from surrounding tissue (see Supplementary Fig. 7) and weighed in wet condition with an accuracy weighing machine (Sartorius Cubis, MSE, Sartorius AG, Göttingen, Germany). Subsequently the radula and jaw (RJ) were separated from the buccal mass musculature (B) manually; RJ and B were weighed in wet condition. B and RJ were then dried for one week and weighed again to obtain data on dry mass. Forces were corrected for dry B to receive Relative Force III (force/ dry B mass, mN/mg) and for dry RJ to receive Relative Force IV (force/dry RJ mass, mN/mg).
For scanning electron microscope (SEM) images radulae and jaws were rewetted and cleaned with proteinase K digesting food particles according to the protocol of 94 , followed by a short ultrasonic bath. Structures were mounted on SEM stubs, coated with palladium and visualized with a Zeiss LEO 1525 (One Zeiss Drive, Thornwood, USA). Radular length, width, and area could be calculated. Forces were corrected for radular area to receive Relative Force V (force/radular area, mN/mm 2 ).

Results
Radula and its teeth. All

Radular motion.
Video footage reveals the radular motion and feeding behaviour (see Supplementary Videos 1 and 2). While feeding, the radula is pushed simultaneously in ventral (vertical down) and anterior (horizontal anterior) direction, before the organ is finally pulled in dorsal (vertical up) and posterior (horizontal posterior) direction and the mouth is closed (see also 87 ). With the first part of the motion the radula loosens food items from the ground and collects particles, which are transported into the mouth opening in the latter phase of the feeding action. When feeding on larger ingesta (e.g. a piece of carrot; see Supplementary Video 1), the anterior part of the radula and the jaw act in concert as counter bearing squeezing and pulling the ingesta. When comparing the feeding behaviours of different individuals we can see that the large sized individuals can completely enclose the carrot piece with their lips, resulting in a dragging on the carrot in vertical direction, tearing large pieces, whereas the small and medium sized individuals are not as comfortable with this due to the small dimension of their mouth. These individuals usually nibble on edges of the item, cutting and slicing smaller pieces in rather horizontal direction employing their radula like a saw, sometimes involving additionally the foot as a clamp (Supplementary Video 2). They also drag the food in ventral direction, but this behaviour is not as pronounced and is not as forcefully as in the smaller specimens (Fig. 4). All gastropods were able to consume the food items offered, but small sized individuals needed to invest approximately 800-900% and medium sized ones 400-500% more time to consume the similar sized food items than large sized gastropods.

Force measurements of body mass classes.
Comparing the Absolute Forces regardless of the direction between the three body mass classes of animals (Fig. 3A) we detect significant differences (p < 0.0001, ChiSquare: 995.6664, df: 2). Large sized individuals are capable of exerting highest forces, followed by the medium sized, and finally small sized individuals (see Table 1). When comparing the Relative Force I regardless of the direction between the three body mass classes of animals (Fig. 3B) we detect significant differences between groups (p < 0.0001, ChiSquare: 409.7297, df: 2). Here, small sized individuals exhibited the highest Relative Force I, followed by the medium sized, and finally large sized animals (see Table 1). Absolute Forces sorted to direction (Fig. 3C) also differed significantly between body mass classes (see Table 1; for statistics see Table 2). For horizontal anterior direction, the highest forces were, however, exerted by the medium sized class, followed by the large, and finally small sized gastropods. For horizontal posterior direction, the highest forces were again exerted by the medium sized class, followed by the small sized gastropods, and finally large ones. In the direction vertical down, the large sized gastropods showed highest forces, followed by www.nature.com/scientificreports/ the medium, and finally small gastropods. The same is found for vertical up: large gastropods, medium, and finally small ones. When comparing the Relative Force I sorted to direction (Fig. 3D) we again detect significant differences between body mass classes (see Table 1; for statistics see Table 2). Here, for horizontal anterior direction, the highest Relative Force I was exerted by the small sized gastropods, followed by the medium, and finally large ones. For horizontal posterior direction, the highest Relative Force I was again exerted by the small, followed by  www.nature.com/scientificreports/ Comparing the Absolute Forces sorted to direction between cohorts we found significant differences ( Fig. 4B; for values see Supplementary Table 2; for statistics see Table 2). For horizontal anterior direction, the highest forces were exerted by H. pomatia, followed by the immature L. fulica, mature L. fulica, A. vulgaris, C. hortensis, and finally C. nemoralis. For horizontal posterior direction, the highest forces were exerted by H. pomatia, followed by C. nemoralis, the immature L. fulica, mature L. fulica, A. vulgaris, and finally C. hortensis. For vertical down direction, the highest forces were exerted by mature L. fulica, followed by H. pomatia, the immature L. fulica, A. vulgaris, C. nemoralis, and finally C. hortensis. For vertical up direction, the highest forces were exerted by mature L. fulica, followed by the immature L. fulica, H. pomatia, C. nemoralis, A. vulgaris, and finally C. hortensis.
When comparing the Relative Force I sorted to direction between cohorts we found significant differences ( Fig. 4C; for values see Supplementary Table 2; for statistics see Table 2). For horizontal anterior, the highest Relative Force I was exerted by C. hortensis, followed by A. vulgaris, the immature L. fulica, C. nemoralis, H. pomatia, and finally mature L. fulica. For horizontal posterior direction, the highest Relative Force I was exerted by C. nemoralis, followed by C. hortensis, A. vulgaris, the immature L. fulica, H. pomatia, and finally  Overall, we found that the body mass (with and without shell) relates in proportion to the masses of the whole buccal mass (wet; BRJ), the radular musculature (buccal mass musculature, wet and dry; B), and the radula and jaw (wet and dry; RJ). When individuals were heavier, they usually possessed higher muscle mass, a heavier radula and jaw (see Supplementary Fig. 8 and Supplementary Table 6). Exceptions were: L. fulica immature no. 9 (18 g body mass and 205.01 mg whole buccal mass) and H. pomatia (38 g body mass and 163.30 mg whole buccal mass), C. nemoralis no. 1 (3.60 g body mass and 27.20 mg whole buccal mass), and A. vulgaris no. 2 (3.50 g body mass and 56.66 mg whole buccal mass). Comparing mature (mature 1: 78.00 g body mass and 286.80 mg BRJ) and immature L. fulica (immature no. 9: 18.00 g body mass and 205.01 mg BRJ) we found that the body mass increases for the factor ~ 4 and BRJ increases for the factor 1.4.
We found that smaller gastropods are capable of exerting higher forces per whole buccal mass, radular muscle, and radula and jaw mass, dry as well as wet, (see Supplementary Fig. 9 and Supplementary Table 6). Cepaea hortensis exerted the highest force per radular musculature mass, followed by C. nemoralis no. 1, A. vulgaris no. 4 and 5, C. nemoralis no. 2, H. pomatia, L. fulica mature no. 1, L. fulica immature no. 9, and finally A. vulgaris no. 1 and 2. The same sequence was also found for force per body mass without shell. With the exception of A. vulgaris no. 4 and 5, exerting the highest forces per radula and jaw mass, we detected the same order for this parameter.
The radular length and width (see Supplementary Fig. 8 and Supplementary Table 5) do not consistently correlate with body mass, the highest radular width was measured for L. fulica mature no Force quotients. We found that the stylommatophoran species perfectly follow the predictions for scaling of force and body mass as the mean Absolute radular Force (regardless of the direction) scales to body mass 0.67 with the quotient 4.25 and the mean Relative Force I (regardless of the direction) scales to body mass −0.33 with the quotient 4.35 (Fig. 5A,B). When the variance of the Absolute Force (regardless of the direction) is scaled to body mass 0.67 and the variance of the Relative Force I (regardless of the direction) is scaled to body mass −0.33 we receive quotients of 3 (Fig. 5C,D). When the forces are sorted to directions the picture is rather puzzling: for horizontal anterior a quotient of 27 (Fig. 6A,B), for horizontal posterior a quotient of 31-32 (Fig. 6C,D), for vertical down a quotient of 2.4 (Fig. 6E,F), and for vertical up a quotient of 0.54 (Fig. 6G,H) is calculated.

Discussion
Force output is often referred to as proportional to muscle mass 0.67 , the muscle cross-sectional area 95,96 , or to body mass 0.6797 whereas the forces, corrected for body mass (force/mass), are referred to be proportional to body mass −0.3397 . The stylommatophoran species examined here, follow the predictions for scaling of force and body mass for mean values. Following previous studies 22, 98 we here experimentally determined a Radular Force Quotient AFQ, AFQ = mean Absolute Force/Bodymass 0.67 , of 4.25 and a Relative Force Quotient RFQ, RFQ = mean Relative Force I/Bodymass −0.33 of 4.35 which can be further used for predictions of forces in Gastropoda. However, when the forces are sorted to directions the picture becomes puzzling, indicating the need for further studies.
The here measured Absolute feeding Forces for Helix pomatia are in a similar range as to those documented by 87 for Cornu aspersum, both gastropod groups have comparable body mass. We detected that gastropods with a higher body mass and a larger body size were capable of exerting higher radular forces, which is not surprising. www.nature.com/scientificreports/ However, the relationship between forces generated by the feeding organ and body size is well documented for vertebrates, but not for molluscs. When forces are corrected for body size or mass, ecological adaptations related with this parameter usually became more pronounced (e.g. 4,7,16,99,100 ; for hypotheses about body size related evolution of bite force, see also 101,102 ). When determining the radular force per body mass (termed Relative Force I), we see that smaller gastropods were capable of exerting the highest forces, followed by the medium, and finally large ones, similar to e.g. carnivorans in placental mammals 7 . This is also not surprising, because of different scaling of body mass and cross-sectional area of muscles 103,104 . Usually, the sampled gastropod specimens exhibiting a higher body mass also possess a proportionally higher mass of buccal mass (correlation between buccal mass size and gastropod body size was also previously described by 51 ). However, the immature and adult Lissachatina fulica, exhibiting strong differences in body mass and size, but having almost similar buccal mass sizes and masses, are an exception to this rule. In most experiments mature L. fulica exerted higher radular forces than the immature gastropods, but some measurements (outliers) revealed that immature L. fulica are also capable of generating the same feeding forces as the mature ones. This could be an indication that the ingesta type does not change during ontogeny; however, this aspect awaits further investigation. But our analyses of Relative Force I (Fig. 4) reveal distinct radular force patterns for cohorts, both Cepaea species have a wide range of exerted forces, followed by Arion, Lissachatina immature, and finally, with the smallest range, Lissachatina mature. This could be an indication that species have species-specific radular forces. Cepaea potentially feeds naturally on a broader food spectrum whereas Lissachatina is more restricted, but this also awaits further investigations. Arion was fed with fresh strawberries, since it did not want to feed on carrots. This could have influenced the forces generated; potentially Arion is capable of exerting higher forces. It should be additionally stated that here the whole body masses were related to the radular forces generated. But, as we experimented with four snail species and one slug species (Arion vulgaris), whole body masses are not directly comparable, because the mass of the shell could cause artefacts. In future studies the body mass without shell should be determined persistently to detect a potentially more accurate relationship or even correlation. In past studies, it has been shown that the forces generated usually correlate with the muscle mass, muscle size, or muscle diameter (for vertebrates: e.g. 8,16,105,106 ; for invertebrates: 107 ). This is congruent to the here observed patterns in gastropods. However, for precise interpretation of the relationship between buccal mass size and the mass of the radular muscles (buccal mass musculature) knowledge about the functional role of each feeding muscle must be available, which is not the case for the taxa studied here. There are some detailed studies analyzing the in vivo buccal mass movement and buccal mass muscle function in Aplysia 89,90,92,108 . But since Aplysia belongs to the Opisthobranchia and is not closely related to the taxa involved in this experiment www.nature.com/scientificreports/ the function of the buccal mass muscles for these stylommatophoran taxa cannot be assigned yet. However, to deeply understand the relationship between measured radular forces involved in specific radular motions, the function of each muscle, their work in concert, as well as the muscle and fiber size, length, or diameter need to be investigated. As there are many different parameters that relate with feeding force (for vertebrates: e.g. the skull geometry and size: 10,17,99,[109][110][111] ; or muscular development: 105 ), it is very difficult to produce reliable models for molluscs due to the lack of solid experimental data. Additionally, studies on invertebrates reveal that muscle stress varies considerably depending on the muscle [112][113][114] . This could also be the case for molluscs as the forces exerted per radular muscular mass differ extremely between the analyzed stylommatophoran individuals (see Supplementary Fig. 9 and Supplementary Table 6). This again shows that pure anatomy-based studies on the muscle systems of invertebrates do not necessarily provide data on physiology (see also 24 ). Additionally, the area of the radula used for foraging could influence results, but the working area could so far only be determined by involving sandpapers of different roughness 115 . Since sandpaper is a rather flat surface in comparison to a slice of carrot we cannot directly translate our past findings to the experiments here.
Our results clearly indicate that larger and heavier animals exerted higher forces in the vertical directions, whereas medium and small individuals exerted higher forces in horizontal directions. This is additionally supported by the analyses of the video footage showing that larger animals rather pull with radula and jaw, whereas smaller individuals use their radular often like a saw in anterior-posterior direction when approaching the food item (see Supplementary Video 1 and 2). This shift in feeding behavior seems to depend on the ingesta size in relation to the mouth opening size. Smaller and medium sized gastropods-even though capable of embracing the whole item with the lip-seem to prefer this alternative feeding pattern. An ingesta-depending shift in feeding pattern, i.e. dynamics of swallowing, had been documented for other gastropod taxa when altering parameters of the ingesta 108 , e.g. its hardness 116 , its load and width 117 , or its size [118][119][120] . Additionally, gastropods of different sizes have been found to feed on different ingesta types, possibly correlating with ontogenetic size changes of the mouthparts 121 .
As already stated above, heavier gastropods were usually capable of exerting higher radular forces than smaller ones (except for immature and adult L. fulica). This indicates that larger gastropods might forage on a broader spectrum of ingesta, as it had been previously described for turtles 109 . However, the smaller individuals were able to exert higher forces per body mass and additionally show a distinct radular motion pattern resulting in a distinct feeding pattern. All experiments resulted in the consumption of the ingesta offered, but smaller gastropods invested more time. This effect had also been observed for lizards 122 and should be further investigated for Mollusca. Thus, we can conclude that in studies of feeding forces of gastropods with a similar radular type, teeth, and radular motion, adaptations to ingesta might only been detected by studying feeding efficiency and time invested. We hope that in the future more biomechanical, physiological and functional morphological studies will approach the topic of trophic specialization in molluscs via feeding force experiments on species possessing distinct radular morphologies and feeding on more types of ingesta.

Data availability
The datasets generated and analysed during the current study (the force measurements) are available from the corresponding author on reasonable request; all other datasets are included in this published article (and its Supplementary Information files). Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.