This page is referenced by:

• 1 2024-09-17T05:59:44+00:00 Annotating FloLo: How the Media Ecology Project Fostered Early Cinema Performance Analysis and Collaborative Research 6 image_header 2024-10-31T03:25:21+00:00

  
  by Jenny Oyallon-Koloski

  University of Illinois at Urbana-Champaign
  
  Abstract

  Introduction
  Annotation Methods
  Annotating Early Cinema Performance
  Early Performance Norms
  Florence Lawrence: Genre Pioneer
  Conclusions
  Additional Bibliography
   

  Introduction

  The Media Ecology Project (MEP) fosters multimodal analysis and scholarly collaboration while providing researchers around the world with a platform to develop innovative analytical methods. Over the past few years, I have had the pleasure of overseeing one of MEP’s pilot projects that focuses on early acting performance norms in films from the Library of Congress Paper Print collection. Our project’s goal was to study what distinguished the performances of Florence Lawrence (known as the “Biograph Girl”) from other performers’ acting modalities during the rise of Hollywood’s star system. MEP contains over thirty short films—made between 1908 and 1913—in which Lawrence appears, digitally scanned from the extensive Paper Print collection. The key research question for this project is what distinguished Lawrence’s performances from other female actors who appeared in films during this period, like Mary Pickford, Blanche Sweet, Dorothy West, Florence La Badie, or Marion Leonard.

  This pilot project used MEP’s digital tools and network of scholars to generate a wealth of annotations documenting the movement patterns of Lawrence and her contemporaries. I guided this collaborative research project by creating a rigid annotation workflow, which is consultable as part of this essay, for analytical consistency and to provide a foundation for future computational analysis. The options available to our team grew out of the specific context of Lawrence’s movement choices and the stylistic constraints of the films in question. However, such an approach is adaptable for those interested in pursuing comparative, close analysis of figure movement or gestures and ideal for a collaborative research process.
  
  Lawrence is considered by some to be the first American film star, and audiences came to know her as the Biograph Girl before film actors were given named credit in films.¹ Our analysis looks at the films Lawrence made while working for Biograph, all directed by D. W. Griffith and filmed between the summers of 1908 and 1909. This is an ideal period to study film style in which figure movement drives narrative comprehension, as these films do not feature any editing within scenes, something that Lea Jacobs and Ben Brewster suggest becomes more dominant in 1913, noting “not only did editing permit the actor to do ‘less’ in terms of posing and gesture, but the pace of a highly edited film required it.”² The films consulted in the Paper Print collection also do not highlight camera movement, meaning that the majority of figure movement through the frame is the result of human agency.³

  Something about Lawrence’s roles and performances caught the public’s attention before film stars were identified by name. So why were audiences so enamored of her? Was Lawrence’s popularity with audiences as the Biograph Girl largely the result of successful marketing strategies? Was there also something unique about her physicality that set her apart? And what can we learn from closer study of these films in this collaborative format about her movement training and influence on early film performance norms? To help answer these questions, we analyzed her gestures and figure movement, using Laban Movement Analysis (LMA) as an analytical framework, with the hope that eventually this granular work could be computerized and codified on a larger scale. Granular analysis, made possible through the combination of figure movement analytics and computational procedures, provided a methodological intervention in the way we study performance in early cinema.

  Annotation Methods

  This collection includes over fifty paper print films starring Lawrence and her contemporaries, all of which are viewable in full through MEP’s online platform. The result is easy accessibility to these films, especially for those scholars who may not have convenient access to on-site holdings of major archives. At the time of annotation creation, the films were viewable through the Mediathread annotation platform, which creates opportunities for pedagogical discussion and collaborative scholarship. Participants can read supplemental materials on the site and contribute to discussion forums. They can also add time-based annotations to the films themselves. Users can generate subclips, as we have done extensively for the Lawrence pilot project, that can include verbal descriptions and metadata, including second-based time increments. These annotations are amplified thanks to a relationship with onomy.org, which provides broad, controlled vocabularies to standardize metadata tags. The Semantic Annotation Tool now allows access to the paper print films as well, adding the additional ability to isolate parts of the frame through a polygon selection in order to clarify which part of the frame the annotation refers to. Because this pilot project began before the completion of the Semantic Annotation Tool, however, we continued the workflow in Mediathread for consistency.

  To study the precise shapes Lawrence and other actors were moving through and how they were blocked in the frame, we needed a more precise vocabulary set tailored to our specific research questions. I amplified our descriptive vocabulary through annotations in Mediathread that drew on the language and structure of LMA, in which I am certified by the Laban/Bartenieff Institute of Movement Studies, New York. For this pilot project, my priorities were the ability to generate annotations of Lawrence’s performance that were systematic in their descriptive precision and create an annotation workflow that would be accessible to participants not trained in LMA. We also wanted to consider computerized methods to analyze the results of our annotations. As a result, the language we generated needed to have a clear syntactic consistency in both semantic sentence structure and vocabulary choice. Given the collaborative nature of our research, we also needed to ensure that scholars would annotate Lawrence’s movement in a standardized way and with the same attention to granular detail. Because of its extensive movement taxonomy, LMA was the ideal methodology to serve as a foundation for this work. 

  LMA and its broader discipline of Laban/Bartenieff Movement Studies provide a descriptive approach to thorough, pattern-based micro- and macroanalysis of the expressivity of human movement. It is, to my knowledge, the most granular and rigorous way to describe, segment, and analyze movement forms. However, this system cannot provide a universal road map for understanding the complexities of human motion; the work, pioneered by Rudolf Laban and his students in the midtwentieth century, was developed from studies of primarily white, European bodies and does not take into account historical or cultural factors that impact the semantics of movement. Rather, the taxonomy at the core of this system —developed by Laban and his students—can describe and analyze movements of the body within a narrow set of parameters that provide a comparative formal baseline from which movement can be contextualized.⁴

  The primary work of an LMA-trained analyst is to discover movement essences: what makes someone’s movement expressive within their unique context. Much of the taxonomy consists of breaking movement down and situating it within various spectra. As Karen Bradley has suggested, “Teasing phenomena into polar opposites and defining continua along the polarities is at the heart of Laban’s work.”⁵ This work also allows for analysis and description of both quantitative and qualitative aspects of movement, which permit the analyst to explore movement from a range of perspectives. For the purposes of computerized data generation, however, much of our scaled analytical work focuses on simple quantitative aspects, as a study of our analytical results will demonstrate. These distinctions are broadly organized into four sections in LMA, each of which seeks to answer a different question. What follows is a summary of the LMA taxonomy.

  First, Space asks Where is the body moving in relation to the environment? This category allows for a more precise way of describing the spatial pulls along which the body and body parts are moving. For the work on Lawrence’s acting style, this category is especially vital to document the spatial placement of her gestures. When actors move their arms, are they more likely to open away from or cross the midline, for example? Do they more frequently extend their limbs far away from them or keep them close to their bodies? What pathways do the gestures tend to employ?⁶

  Second, Body asks What parts of the body are moving and in what sequence? The Body category provides useful guidelines related to the various body parts and joints and encourages the analyst to reflect on various anatomical and neuromuscular considerations. This work also provides a vocabulary for the basic actions of the body, which are particularly useful for categorizing the ways that Lawrence moves through the frame. These Basic Body Actions—traveling, change of support, rotation, flexion, extension, jumping, and falling—provide a basis for our staging annotations.

  Third, Effort asks How is the body moving? Laban suggests that all dynamic changes in the body can be broken down into four motion factors—Space, Weight, Time, and Flow—and these motion factors can fluctuate between two polar opposite qualities, also called Effort elements. Those Effort elements can manifest individually or in combination to create greater or lesser intensity.⁷

  Finally, Shape asks How is the body relating to its environment? A study of Shape encourages the analyst to consider the movers in their spatial context in order to understand their motivations relative to the surrounding environment.⁸

  The Effort and Space categories draw on complex theoretical frameworks that require extensive study to apply consistently. As a result, we used them less heavily for our annotations. However, simple qualitative distinctions, related to Time Effort especially, do work their way into some of our annotations. For example, is Lawrence walking (a slower, more sustained pace) or running (a quicker pace)? Does she take a step (which often requires a lighter Weight Effort) or jump (which requires a greater increase of Strong Weight)? We can also have a sense of the quickness of certain gestures based on the average lengths of the subclips generated in Mediathread. 

  Body-Effort-Space-Shape provides the foundation of the LMA tool kit. While it cannot serve as a substitution for historical knowledge of a particular movement form, the thoroughness of the taxonomy allows the analyst to see patterns that are present in seemingly disparate movement practices. The precise vocabulary of the Laban system can make it less accessible to people who are not thoroughly versed in the approach. However, the specificity of language makes this movement taxonomy particularly beneficial for granular precision and the elimination of synonyms that could impede the analysis of computer-based annotations and metadata as well as machine vision protocols. Inherent in the work of a Laban Movement analyst is the acknowledgment of and attempt to overcome personal bias. Absolute objectivity is not attainable, and individual areas of interest and personal predilections in relation to movement analysis and screen acting are visible in the differing research programs of each scholar contributing to the Lawrence pilot project. However, the precision of LMA’s descriptive language, especially in the more quantitative categories of Body and Space, allowed us to generate a high volume of granular analysis regarding Lawrence’s acting choices on-screen. 

  In addition to serving as a research space in our study of these films, the Mediathread platform also exists as a place to workshop and discuss methodological approaches. The choice to use LMA as an analytical tool for this project was a topic that several scholars debated in a Mediathread discussion board created by Elisa Uffreduzzi titled “Laban Movement Analysis and other theories of body movement . . . or not?” Other Mediathread assets include its facilitation of scholarly discussions and its ability to record and archive the progress of collaborative work such as ours.

  LMA’s broad applicability makes it valuable for an array of movement analysis–based projects. Another factor was the smooth integration of this analytical approach with existing moving image analytical work, of staging especially. The current analytical focus of MEP’s pilot project is Lawrence’s movement baseline, but our descriptive strategies can be beneficial to an exploration of any actor’s performance. Indeed, we developed our annotation structure with comparison in mind so that existing tools could serve, through additional analysis of silent actors’ movement, to situate Lawrence in context with her peers. Various collaborators working on this project were therefore able to apply these same annotation guidelines to other performers, like Pickford, who feature frequently in the Paper Print collection.

  The result of this annotation labor is a wealth of information about the basic staging and acting motions of Lawrence, Pickford, and other performers. Numerous film scholars contributed to the labor-intensive generation of this invaluable data, including Mark Williams (Dartmouth College), Uffreduzzi (University of Rome), Danae Kleida (Utrecht University), and a number of Williams’s Dartmouth students: Brittany Murphy, Nora Plant, Long Do, Anna Glasgow, Ileana Sung, and Alexandra Salyer. Many of their annotative contributions followed the rigid guidelines I created, adding to our collective pool of knowledge about fundamental staging and acting choices, while the flexibility of the Mediathread annotation platform also allowed our collaborators to explore specific acting conventions for their own research purposes as well.

  The result is a consistency of descriptive precision across the annotations generated on Lawrence’s and her contemporaries’ performance modalities. This foundational data allows us to compare her performances to those of other actors and apply consistent analytical metrics to her performances in different genres. 

  Annotating Early Cinema Performance

  Annotation priorities were documenting first the staging of the performer in the frame and second the frequency and types of gestures. In our work on the first of these categories, we sought to understand the sort of movement the actor employed to get from Point A to Point B, where she was in the space, and how frequently she moved (the duration of each new shift in space or stationary pause). As a result, each annotation contains a descriptive sentence, written with a consistent semantic structure, along with metadata that documents actors’ positions in the frame and the general action they employ to move from Point A to Point B. The action verbs I chose derive in large part from the Basic Body Actions described in LMA. Metadata and descriptions also track position in the frame, segmented according to depth, horizontality, and occasionally verticality. The vocabulary and underlying principles used to study cinematic staging guide our annotative methodology for this work, as articulated in Jacobs and Brewster’s From Theatre to Cinema. These annotations document the moments in which the performer remains stationary as well. This proved to be an important variable and allowed us to isolate when a character remained in a portion of the frame for an extended period. Figure 1 demonstrates how this annotative work appears in the Mediathread platform. The full guidelines can be found here. This document provides the guidelines used in this MEP project to annotate early cinema performers’ staging and blocking.

  The second category of annotations was more individualized. I developed metrics based predominantly on the Space category of LMA to document the various gestures Lawrence executes with granular precision. In this category of annotations, I sought to learn the kinds of isolated and full-body gestures that Lawrence employs, the frequency of her gestures, and the dominant body parts or types of gestures she relies upon. The resulting metadata and descriptions use the Space vocabulary from LMA to indicate where the body parts are moving in space—in relation to horizontality, verticality, and depth—the general pathway of the gesture, and the isolation of the specific body part moving. This work allowed me to document how frequently actors use full-body actions as opposed to isolated gestures, the body parts they most frequently gesture with (one arm, both arms, etc.), and what portion of the space within their reach—what Laban called the “kinesphere”—they most frequently move in.⁹

  Observation of Lawrence and her contemporaries demonstrates that many of their movements, head and arm gestures especially, last for less than a second and occasionally only a few frames. One of Mediathread’s technological limitations is the inability to create subclips lasting less than a second. Because of this, the platform was not able to record the duration of these gestures with great precision. Any determinations of gestural patterns in Mediathread, therefore, cannot be isolated precisely without frame-based temporal adjustments. In addition, comprehensive movement description that requires frame-level annotations demands a significant time investment. For these reasons as well as the others discussed, our annotation efforts focused on the staging and more holistic movements of the performers through the mise-en-scène. However, Figure 2 demonstrates how this annotative work on gestures appears in the Mediathread platform. Some of our research collaborators, Kleida in particular, also used their own methods to study these films’ gestural patterns.¹⁰
  

  Early Performance Norms

  This collective annotative metadata allows us to better confirm a few crucial staging and acting norms from the films in the Paper Print collection. The total annotations, all coded individually, provide approximately 2,600 action metadata tags and 7,000 staging and blocking metadata tags. From this information we can identify several broad trends. We can observe where in the frame these performers are most likely to appear (foreground, midground, or background, right, left, or center). We can see how often performers are shown entering or exiting the frame. And we can categorize the basic acting gestures they most often employ while moving through the mise-en-scène or as they stand in place. Table 1 summarizes these findings. 

   

  Table 1. Metadata Staging and General Movement Results from Mediathread Annotations
  

  Staging position (Lawrence)
  Sagittal movement (z-axis)
  FG	617
  MG	606
  BG	135
  Bgoff	20
  Horizontal movement (x-axis)
  RightOff	54
  Right	244
  RightCenter	204
  Center	341
  LeftCenter	245
  Left	248
  LeftOff	56
  Vertical movement (y-axis)
  High	13
  Low	36

   

  Staging position (all performers)
  Sagittal movement (z-axis)
  FG	2003
  MG	848
  BG	320
  Bgoff	50
  Horizontal movement (x-axis)
  RightOff	120
  Right	586
  RightCenter	551
  Center	896
  LeftCenter	770
  Left	603
  LeftOff	132
  Vertical movement (y-axis)
  High	42
  Low	71

   

  Broad action (Lawrence)	Totals
  Locomotion (walks, walks backward/upstairs/downstairs, steps, runs)	553 (53%)
  Postural/in place (crouches, stands, sits down, stands up)	464 (45%)
  Dramatic movements (crawls, rolls, jumps, falls down)	23 (2%)

   

  Broad action (all performers)	Totals
  Locomotion (walks, walks backward/upstairs/downstairs, steps, runs, is moved)	1,392 (54.4%)
  Postural/in place (crouches, stands, is held, rotates, sits down, stands up)	1,132 (44.2%)
  Dramatic movements (crawls, rolls, jumps, falls down)	36 (1.4%)

  The annotation data comes from a range of lead performers in these films, notably Lawrence, Pickford, Sweet, West, La Badie, and Leonard. The broad patterns in staging reinforce the importance of placing the lead actors in the foreground for visibility, as these performers consistently spend more time in the foreground than they do in the midground and background combined. The horizontal distribution shows that the staging in these films made use of the full horizontal space of the frame, but also that these performers were frequently shown exiting the mise-en-scène, likely to use the movement of the performers as a way of concluding a scene or as a match on action with the subsequent tableau setting. The broader actions of the performers are largely divided between actions that hold them in a stationary part of the frame—standing, sitting, rotating, crouching—and those that show them moving through the environment—running, walking, taking a step, or walking up and down stairs. Few of these performance moments, however, engage in more dramatic movements like rolling, falling, jumping, or crawling.

  Metadata specific to Lawrence demonstrates the versatility of her performances and the clear stylistic choices made to draw attention to her in the frame. In a similar fashion to the general data about staging positions, she is more frequently in the foreground and midground, as placement in the background would make her less visible. She also appears more often in the right or left of the frame than she does in the center, which follows established composition theories that posit viewers are more likely to look at information placed in the outer thirds of the frame. The data also shows that Lawrence, like all these performers generally, appears more frequently in the left side of the frame than the right, an interesting observation that merits additional inquiry.

  Lawrence’s broader performance strategies, however, differentiate her from the other performers analyzed here (see Table 2). She is frequently still in the frame, holding her staging position, which is an effective way to keep attention on herself when other actors are moving around her. However, unlike the more balanced split between stillness and motion in the group of performances, Lawrence is more frequently moving through the frame. She changes position frequently, most often by walking (taking multiple steps at an average pace) or taking a single step. In addition, she often uses her body for comedic effect. These simple movement strategies are punctuated by quicker or more humorous movement through the space, in the more comedic Jones series especially. Lawrence crouches down and crawls across the ground in one performance (Mrs. Jones’ Lover [Griffith, 1909]) and falls down in another (Mr. Jones at the Ball [Griffith, 1908]).
  Table 2. Metadata Action Verb Results from Mediathread Annotations Titled “Florence Lawrence Movement”
  

  Action metadata (Lawrence)
  Action	Metadata	Totals (23 films)
  Walks (takes multiple steps at an average pace)	walks	292
  Stands (remains in one place)	still	179
  Stands up	stands	178
  Steps (takes a single step)	steps	161
  Sits down	sits	70
  Walks/steps backward (a qualifier to be added after the action verb)	backward	66
  Crouches (moves to a lower position, creating a ball shape)	crouches	37
  Runs (takes multiple steps at a faster pace)	runs	28
  Falls down (to the ground)	falls	15
  Walks up stairs	stairsup	3
  Walks down stairs	stairsdown	3
  Rolls (rotates the body horizontally)	rolls	3
  Jumps	jumps	3
  Crawls (moves on hands and knees)	crawls	2
  Is held (physically restrained)	held	0

  Lawrence frequently also draws attention to her face and facial expressions through her hand gestures, as seen in Figure 1. This performance pattern is observable even though it is an area of inquiry that could be further quantified with Mediathread annotations. Lawrence often brings a hand or both hands to her face, which encourages the audience’s eye to move there as well. Some of these gestures are more codified as indicative of a particular emotion: pressing hands to cheeks to convey shock or holding her head with her hands to convey sorrow. Many others, however, are small, idle gestures—like touching her hair or scratching her neck—that guide the viewer’s eye to Lawrence’s more subtle expressions. Analytical observations like these are easy to study and annotate in Mediathread thanks to the ability to generate multiple annotations for the same time code, allowing us to explore and document such micropatterns of movement without sacrificing existing work.

  The most striking observation from this collective metadata is the source of the majority of the instances of performers rolling, crawling, jumping, or running. Most come from Lawrence’s films. This data points to her performative range as well as to the more energetic and comedic aspects of her acting choices, ones that distinguish her from her female contemporaries.

  Florence Lawrence: Genre Pioneer

  The data gathered from these annotations points to Lawrence’s performative versatility. Further study of her films also demonstrates how she created her own lexicon of gestures and meaningful postures, setting performative expectations for her audiences in the melodramatic and comedic genre films she most often starred in.

  Lawrence likely trained in a range of popular acting and movement methods of the period, American Delsartism in particular.¹¹ Delsarte’s name quickly became an umbrella term for physical practices offering lessons in “expression,” both for training in the professional trades of theater and dance and recreationally as a form of exercise that would promote cultural knowledge, health, and beauty. These approaches in the US were predominantly pitched at white, middle- and upper-class women and tended to emphasize the upper body through gesture and draw attention away from the lower half of the body, especially the legs, which were often hidden under floor-length skirts.¹² The gestural emphasis of Delsartean movement, therefore, was closely tied to American perceptions of moral decency. 

  We can observe glimmers of Delsartean influence in Lawrence’s performances, notably the graceful, off-balance, oppositional qualities associated with the movement practice. The shapes of her curved wrist in several of these moments stand out, particularly in relation to sketches of Delsartean poses in primary documents. We can compare, for example, Lawrence’s suspended fall in The Salvation Army Lass (Figure 5) to Morgan’s sketch of the Delsartean point indicating “Go!” (Figure 6).¹³ Notably, examples of Delsartean aesthetics come from both Lawrence’s comedic and dramatic roles. These moments also convey less obvious emotional expression. They are more complex and graceful than some of the previous examples, but also more narratively opaque. For the most part, Lawrence’s emotional performances do not align with existing examples of the Delsartean lexicon. More often we can see her creating her own taxonomy of gestural and postural movements, which she repeats across her film corpus, to embody clear emotional moments like shock and horror, despair, grief, love-derived joy, and so on. This pattern of repetition is especially clear in her moments of jealous anger, a common emotion for her character Mrs. Jones in the comedy series (Video 1). These examples have greater variation in her physical form but ultimately all lead to the same general action: bouncing in anger before hitting the object of her jealousy—usually John Cumpson, who plays Mr. Jones. The rapidity and largely ungraceful nature of Lawrence’s movements are particularly comedic and un-Delsartean. The similarities in Mrs. Jones’s fits of violence against her husband here are also clearly an intentional choice, serving as an expected comical convention of the films.

  That Lawrence repeats movements to express similar emotions is not surprising, given that she was often acting in two releases a week while at Biograph and genres are predicated on formulaic repeatability.¹⁴ She demonstrates a nimble skillfulness in her ability to meld grace with conventional gestures and a rougher slapstick physicality based on the varied needs of these early Biograph plots. 

  Conclusions

  The “FloLo” pilot project for MEP and its analysis of Lawrence’s performance modalities reveal her to be a versatile actor with an expansive physicality who quickly understood the value of generic repeatability. She displays a keen ability to hold the viewer’s attention as she moves through the frame and is able to shape performances across a range of film genres, comedy and melodrama in particular. The use of the movement taxonomy from LMA as the foundation for my analytical methodology allows for a more objective annotation structure that encourages descriptive consistency among our annotation contributors. The broad applicability of this system makes this approach to studying actors’ movement beneficial to research on all performers, not just Lawrence. Moreover, this project demonstrates the utility of MEP’s integration with the Mediathread and Semantic Annotation Tool platforms for collaborative data generation, outcome analysis, and dissemination of data to other formats for further visualization. 

  Our research findings also point to additional questions: how innovative were Lawrence’s staging conventions compared to those of her contemporaries or the norms of legitimate theater? Do other actors of the period share Lawrence’s physical versatility and employ the same gestural patterns to draw attention to themselves? How can this research approach be expanded to study silent film actresses in other filmmaking traditions? And how can we improve our annotation and computerized methods to effectively and efficiently answer these questions? The Lawrence pilot project created opportunities to share information and develop collaborative, networked scholarship. These productive questions will continue to benefit from additional interdisciplinary inquiry and scholarly collaboration.

  Additional Bibliography

  Fowle Adams, Florence. Gesture and Pantomimic Action, 2nd ed. New York: Edgar S. Werner, 1891.
  Hennefeld, Maggie. Specters of Slapstick and Silent Film Comediennes. New York: Columbia University Press, 2018.
  Mayer, David. “Deep Theatrical Roots: Griffith and the Theater.” In A Companion to D. W. Griffith, edited by Charlie Keil. John Wiley & Sons, 2018.
  Preston, Carrie J. Modernism’s Mythic Pose: Gender, Genre, Solo Performance. Oxford: Oxford University Press, 2011.
  Stebbins, Genevieve. Delsarte System of Expression. New York: Edgar S. Werner, 1887.

  A list of external links featured in this essay can be found here.¹⁵
  
  Jenny Oyallon-Koloski is an assistant professor of media and cinema studies at the University of Illinois at Urbana-Champaign. Her first book, Storytelling in Motion: Cinematic choreography and the film musical, is forthcoming from Oxford University Press. She also co-directs the movement visualization lab (http://mvlab.org/) and serves as a movement analysis specialist for the Media Ecology Project.
  
  Title Image: Mr. Jones at the Ball (Griffith, 1909), scan provided by Film Preservation Society

  Annotating FloLo: How the Media Ecology Project Fostered Early Cinema Performance Analysis and Collaborative Research © 2024 by Jenny Oyallon-Koloski is licensed under CC BY 4.0.
  
  The Journal of e-Media Studies is published by the Dartmouth Library.